---
title: Getting Started with Datadog
description: Datadog, the leading service for cloud-scale monitoring.
breadcrumbs: Docs > Infrastructure > Datadog Resource Catalog
---

# gcp_bigquery_model{% #gcp_bigquery_model %}

## `ancestors`{% #ancestors %}

**Type**: `UNORDERED_LIST_STRING`

## `best_trial_id`{% #best_trial_id %}

**Type**: `INT64`**Provider name**: `bestTrialId`**Description**: The best trial_id across all training runs.

## `creation_time`{% #creation_time %}

**Type**: `INT64`**Provider name**: `creationTime`**Description**: Output only. The time when this model was created, in millisecs since the epoch.

## `default_trial_id`{% #default_trial_id %}

**Type**: `INT64`**Provider name**: `defaultTrialId`**Description**: Output only. The default trial_id to use in TVFs when the trial_id is not passed in. For single-objective [hyperparameter tuning](https://docs.datadoghq.com/bigquery-ml/docs/reference/standard-sql/bigqueryml-syntax-hp-tuning-overview) models, this is the best trial ID. For multi-objective [hyperparameter tuning](https://docs.datadoghq.com/bigquery-ml/docs/reference/standard-sql/bigqueryml-syntax-hp-tuning-overview) models, this is the smallest trial ID among all Pareto optimal trials.

## `description`{% #description %}

**Type**: `STRING`**Provider name**: `description`**Description**: Optional. A user-friendly description of this model.

## `encryption_configuration`{% #encryption_configuration %}

**Type**: `STRUCT`**Provider name**: `encryptionConfiguration`**Description**: Custom encryption configuration (e.g., Cloud KMS keys). This shows the encryption configuration of the model data while stored in BigQuery storage. This field can be used with PatchModel to update encryption key for an already encrypted model.

- `kms_key_name`**Type**: `STRING`**Provider name**: `kmsKeyName`**Description**: Optional. Describes the Cloud KMS encryption key that will be used to protect destination BigQuery table. The BigQuery Service Account associated with your project requires access to this encryption key.

## `etag`{% #etag %}

**Type**: `STRING`**Provider name**: `etag`**Description**: Output only. A hash of this resource.

## `expiration_time`{% #expiration_time %}

**Type**: `INT64`**Provider name**: `expirationTime`**Description**: Optional. The time when this model expires, in milliseconds since the epoch. If not present, the model will persist indefinitely. Expired models will be deleted and their storage reclaimed. The defaultTableExpirationMs property of the encapsulating dataset can be used to set a default expirationTime on newly created models.

## `feature_columns`{% #feature_columns %}

**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `featureColumns`**Description**: Output only. Input feature columns that were used to train this model.

- `name`**Type**: `STRING`**Provider name**: `name`**Description**: Optional. The name of this field. Can be absent for struct fields.
- `type`**Type**: `STRUCT`**Provider name**: `type`**Description**: Optional. The type of this parameter. Absent if not explicitly specified (e.g., CREATE FUNCTION statement can omit the return type; in this case the output parameter does not have this "type" field).
  - `struct_type`**Type**: `STRUCT`**Provider name**: `structType`**Description**: The fields of this struct, in order, if type_kind = "STRUCT".

  - `type_kind`**Type**: `STRING`**Provider name**: `typeKind`**Description**: Required. The top level type of this field. Can be any GoogleSQL data type (e.g., "INT64", "DATE", "ARRAY").**Possible values**:

    - `TYPE_KIND_UNSPECIFIED` - Invalid type.
    - `INT64` - Encoded as a string in decimal format.
    - `BOOL` - Encoded as a boolean 'false' or 'true'.
    - `FLOAT64` - Encoded as a number, or string 'NaN
    - `STRING` - Infinity' or '-Infinity'.
    - `BYTES` - Encoded as a string value.
    - `TIMESTAMP` - Encoded as a base64 string per RFC 4648, section 4.
    - `DATE` - Encoded as an RFC 3339 timestamp with mandatory 'Z' time zone string: 1985-04-12T23:20:50.52Z
    - `TIME` - Encoded as RFC 3339 full-date format string: 1985-04-12
    - `DATETIME` - Encoded as RFC 3339 partial-time format string: 23:20:50.52
    - `INTERVAL` - Encoded as RFC 3339 full-date 'T' partial-time: 1985-04-12T23:20:50.52
    - `GEOGRAPHY` - Encoded as fully qualified 3 part: 0-5 15 2:30:45.6
    - `NUMERIC` - Encoded as WKT
    - `BIGNUMERIC` - Encoded as a decimal string.
    - `JSON` - Encoded as a decimal string.
    - `ARRAY` - Encoded as a string.
    - `STRUCT` - Encoded as a list with types matching Type.array_type.

## `friendly_name`{% #friendly_name %}

**Type**: `STRING`**Provider name**: `friendlyName`**Description**: Optional. A descriptive name for this model.

## `hparam_search_spaces`{% #hparam_search_spaces %}

**Type**: `STRUCT`**Provider name**: `hparamSearchSpaces`**Description**: Output only. All hyperparameter search spaces in this model.

- `activation_fn`**Type**: `STRUCT`**Provider name**: `activationFn`**Description**: Activation functions of neural network models.
  - `candidates`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `candidates`**Description**: Canididates for the string or enum parameter in lower case.
- `batch_size`**Type**: `STRUCT`**Provider name**: `batchSize`**Description**: Mini batch sample size.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the int hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `candidates`**Description**: Candidates for the int parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the int hyperparameter.
    - `max`**Type**: `INT64`**Provider name**: `max`**Description**: Max value of the int parameter.
    - `min`**Type**: `INT64`**Provider name**: `min`**Description**: Min value of the int parameter.
- `booster_type`**Type**: `STRUCT`**Provider name**: `boosterType`**Description**: Booster type for boosted tree models.
  - `candidates`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `candidates`**Description**: Canididates for the string or enum parameter in lower case.
- `colsample_bylevel`**Type**: `STRUCT`**Provider name**: `colsampleBylevel`**Description**: Subsample ratio of columns for each level for boosted tree models.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `colsample_bynode`**Type**: `STRUCT`**Provider name**: `colsampleBynode`**Description**: Subsample ratio of columns for each node(split) for boosted tree models.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `colsample_bytree`**Type**: `STRUCT`**Provider name**: `colsampleBytree`**Description**: Subsample ratio of columns when constructing each tree for boosted tree models.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `dart_normalize_type`**Type**: `STRUCT`**Provider name**: `dartNormalizeType`**Description**: Dart normalization type for boosted tree models.
  - `candidates`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `candidates`**Description**: Canididates for the string or enum parameter in lower case.
- `dropout`**Type**: `STRUCT`**Provider name**: `dropout`**Description**: Dropout probability for dnn model training and boosted tree models using dart booster.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `hidden_units`**Type**: `STRUCT`**Provider name**: `hiddenUnits`**Description**: Hidden units for neural network models.
  - `candidates`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `candidates`**Description**: Candidates for the int array parameter.
    - `elements`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `elements`**Description**: Elements in the int array.
- `l1_reg`**Type**: `STRUCT`**Provider name**: `l1Reg`**Description**: L1 regularization coefficient.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `l2_reg`**Type**: `STRUCT`**Provider name**: `l2Reg`**Description**: L2 regularization coefficient.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `learn_rate`**Type**: `STRUCT`**Provider name**: `learnRate`**Description**: Learning rate of training jobs.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `max_tree_depth`**Type**: `STRUCT`**Provider name**: `maxTreeDepth`**Description**: Maximum depth of a tree for boosted tree models.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the int hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `candidates`**Description**: Candidates for the int parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the int hyperparameter.
    - `max`**Type**: `INT64`**Provider name**: `max`**Description**: Max value of the int parameter.
    - `min`**Type**: `INT64`**Provider name**: `min`**Description**: Min value of the int parameter.
- `min_split_loss`**Type**: `STRUCT`**Provider name**: `minSplitLoss`**Description**: Minimum split loss for boosted tree models.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `min_tree_child_weight`**Type**: `STRUCT`**Provider name**: `minTreeChildWeight`**Description**: Minimum sum of instance weight needed in a child for boosted tree models.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the int hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `candidates`**Description**: Candidates for the int parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the int hyperparameter.
    - `max`**Type**: `INT64`**Provider name**: `max`**Description**: Max value of the int parameter.
    - `min`**Type**: `INT64`**Provider name**: `min`**Description**: Min value of the int parameter.
- `num_clusters`**Type**: `STRUCT`**Provider name**: `numClusters`**Description**: Number of clusters for k-means.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the int hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `candidates`**Description**: Candidates for the int parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the int hyperparameter.
    - `max`**Type**: `INT64`**Provider name**: `max`**Description**: Max value of the int parameter.
    - `min`**Type**: `INT64`**Provider name**: `min`**Description**: Min value of the int parameter.
- `num_factors`**Type**: `STRUCT`**Provider name**: `numFactors`**Description**: Number of latent factors to train on.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the int hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `candidates`**Description**: Candidates for the int parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the int hyperparameter.
    - `max`**Type**: `INT64`**Provider name**: `max`**Description**: Max value of the int parameter.
    - `min`**Type**: `INT64`**Provider name**: `min`**Description**: Min value of the int parameter.
- `num_parallel_tree`**Type**: `STRUCT`**Provider name**: `numParallelTree`**Description**: Number of parallel trees for boosted tree models.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the int hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `candidates`**Description**: Candidates for the int parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the int hyperparameter.
    - `max`**Type**: `INT64`**Provider name**: `max`**Description**: Max value of the int parameter.
    - `min`**Type**: `INT64`**Provider name**: `min`**Description**: Min value of the int parameter.
- `optimizer`**Type**: `STRUCT`**Provider name**: `optimizer`**Description**: Optimizer of TF models.
  - `candidates`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `candidates`**Description**: Canididates for the string or enum parameter in lower case.
- `subsample`**Type**: `STRUCT`**Provider name**: `subsample`**Description**: Subsample the training data to grow tree to prevent overfitting for boosted tree models.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.
- `tree_method`**Type**: `STRUCT`**Provider name**: `treeMethod`**Description**: Tree construction algorithm for boosted tree models.
  - `candidates`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `candidates`**Description**: Canididates for the string or enum parameter in lower case.
- `wals_alpha`**Type**: `STRUCT`**Provider name**: `walsAlpha`**Description**: Hyperparameter for matrix factoration when implicit feedback type is specified.
  - `candidates`**Type**: `STRUCT`**Provider name**: `candidates`**Description**: Candidates of the double hyperparameter.
    - `candidates`**Type**: `UNORDERED_LIST_DOUBLE`**Provider name**: `candidates`**Description**: Candidates for the double parameter in increasing order.
  - `range`**Type**: `STRUCT`**Provider name**: `range`**Description**: Range of the double hyperparameter.
    - `max`**Type**: `DOUBLE`**Provider name**: `max`**Description**: Max value of the double parameter.
    - `min`**Type**: `DOUBLE`**Provider name**: `min`**Description**: Min value of the double parameter.

## `hparam_trials`{% #hparam_trials %}

**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `hparamTrials`**Description**: Output only. Trials of a [hyperparameter tuning](https://docs.datadoghq.com/bigquery-ml/docs/reference/standard-sql/bigqueryml-syntax-hp-tuning-overview) model sorted by trial_id.

- `end_time_ms`**Type**: `INT64`**Provider name**: `endTimeMs`**Description**: Ending time of the trial.
- `error_message`**Type**: `STRING`**Provider name**: `errorMessage`**Description**: Error message for FAILED and INFEASIBLE trial.
- `eval_loss`**Type**: `DOUBLE`**Provider name**: `evalLoss`**Description**: Loss computed on the eval data at the end of trial.
- `evaluation_metrics`**Type**: `STRUCT`**Provider name**: `evaluationMetrics`**Description**: Evaluation metrics of this trial calculated on the test data. Empty in Job API.
  - `arima_forecasting_metrics`**Type**: `STRUCT`**Provider name**: `arimaForecastingMetrics`**Description**: Populated for ARIMA models.
    - `arima_fitting_metrics`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `arimaFittingMetrics`**Description**: Arima model fitting metrics.
      - `aic`**Type**: `DOUBLE`**Provider name**: `aic`**Description**: AIC.
      - `log_likelihood`**Type**: `DOUBLE`**Provider name**: `logLikelihood`**Description**: Log-likelihood.
      - `variance`**Type**: `DOUBLE`**Provider name**: `variance`**Description**: Variance.
    - `arima_single_model_forecasting_metrics`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `arimaSingleModelForecastingMetrics`**Description**: Repeated as there can be many metric sets (one for each model) in auto-arima and the large-scale case.
      - `arima_fitting_metrics`**Type**: `STRUCT`**Provider name**: `arimaFittingMetrics`**Description**: Arima fitting metrics.
        - `aic`**Type**: `DOUBLE`**Provider name**: `aic`**Description**: AIC.
        - `log_likelihood`**Type**: `DOUBLE`**Provider name**: `logLikelihood`**Description**: Log-likelihood.
        - `variance`**Type**: `DOUBLE`**Provider name**: `variance`**Description**: Variance.
      - `has_drift`**Type**: `BOOLEAN`**Provider name**: `hasDrift`**Description**: Is arima model fitted with drift or not. It is always false when d is not 1.
      - `has_holiday_effect`**Type**: `BOOLEAN`**Provider name**: `hasHolidayEffect`**Description**: If true, holiday_effect is a part of time series decomposition result.
      - `has_spikes_and_dips`**Type**: `BOOLEAN`**Provider name**: `hasSpikesAndDips`**Description**: If true, spikes_and_dips is a part of time series decomposition result.
      - `has_step_changes`**Type**: `BOOLEAN`**Provider name**: `hasStepChanges`**Description**: If true, step_changes is a part of time series decomposition result.
      - `non_seasonal_order`**Type**: `STRUCT`**Provider name**: `nonSeasonalOrder`**Description**: Non-seasonal order.
        - `d`**Type**: `INT64`**Provider name**: `d`**Description**: Order of the differencing part.
        - `p`**Type**: `INT64`**Provider name**: `p`**Description**: Order of the autoregressive part.
        - `q`**Type**: `INT64`**Provider name**: `q`**Description**: Order of the moving-average part.
      - `seasonal_periods`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `seasonalPeriods`**Description**: Seasonal periods. Repeated because multiple periods are supported for one time series.
      - `time_series_id`**Type**: `STRING`**Provider name**: `timeSeriesId`**Description**: The time_series_id value for this time series. It will be one of the unique values from the time_series_id_column specified during ARIMA model training. Only present when time_series_id_column training option was used.
      - `time_series_ids`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `timeSeriesIds`**Description**: The tuple of time_series_ids identifying this time series. It will be one of the unique tuples of values present in the time_series_id_columns specified during ARIMA model training. Only present when time_series_id_columns training option was used and the order of values here are same as the order of time_series_id_columns.
    - `has_drift`**Type**: `UNORDERED_LIST_BOOLEAN`**Provider name**: `hasDrift`**Description**: Whether Arima model fitted with drift or not. It is always false when d is not 1.
    - `non_seasonal_order`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `nonSeasonalOrder`**Description**: Non-seasonal order.
      - `d`**Type**: `INT64`**Provider name**: `d`**Description**: Order of the differencing part.
      - `p`**Type**: `INT64`**Provider name**: `p`**Description**: Order of the autoregressive part.
      - `q`**Type**: `INT64`**Provider name**: `q`**Description**: Order of the moving-average part.
    - `seasonal_periods`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `seasonalPeriods`**Description**: Seasonal periods. Repeated because multiple periods are supported for one time series.
    - `time_series_id`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `timeSeriesId`**Description**: Id to differentiate different time series for the large-scale case.
  - `binary_classification_metrics`**Type**: `STRUCT`**Provider name**: `binaryClassificationMetrics`**Description**: Populated for binary classification/classifier models.
    - `aggregate_classification_metrics`**Type**: `STRUCT`**Provider name**: `aggregateClassificationMetrics`**Description**: Aggregate classification metrics.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: Accuracy is the fraction of predictions given the correct label. For multiclass this is a micro-averaged metric.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The F1 score is an average of recall and precision. For multiclass this is a macro-averaged metric.
      - `log_loss`**Type**: `DOUBLE`**Provider name**: `logLoss`**Description**: Logarithmic Loss. For multiclass this is a macro-averaged metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: Precision is the fraction of actual positive predictions that had positive actual labels. For multiclass this is a macro-averaged metric treating each class as a binary classifier.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: Recall is the fraction of actual positive labels that were given a positive prediction. For multiclass this is a macro-averaged metric.
      - `roc_auc`**Type**: `DOUBLE`**Provider name**: `rocAuc`**Description**: Area Under a ROC Curve. For multiclass this is a macro-averaged metric.
      - `threshold`**Type**: `DOUBLE`**Provider name**: `threshold`**Description**: Threshold at which the metrics are computed. For binary classification models this is the positive class threshold. For multi-class classfication models this is the confidence threshold.
    - `binary_confusion_matrix_list`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `binaryConfusionMatrixList`**Description**: Binary confusion matrix at multiple thresholds.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: The fraction of predictions given the correct label.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The equally weighted average of recall and precision.
      - `false_negatives`**Type**: `INT64`**Provider name**: `falseNegatives`**Description**: Number of false samples predicted as false.
      - `false_positives`**Type**: `INT64`**Provider name**: `falsePositives`**Description**: Number of false samples predicted as true.
      - `positive_class_threshold`**Type**: `DOUBLE`**Provider name**: `positiveClassThreshold`**Description**: Threshold value used when computing each of the following metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: The fraction of actual positive predictions that had positive actual labels.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: The fraction of actual positive labels that were given a positive prediction.
      - `true_negatives`**Type**: `INT64`**Provider name**: `trueNegatives`**Description**: Number of true samples predicted as false.
      - `true_positives`**Type**: `INT64`**Provider name**: `truePositives`**Description**: Number of true samples predicted as true.
    - `negative_label`**Type**: `STRING`**Provider name**: `negativeLabel`**Description**: Label representing the negative class.
    - `positive_label`**Type**: `STRING`**Provider name**: `positiveLabel`**Description**: Label representing the positive class.
  - `clustering_metrics`**Type**: `STRUCT`**Provider name**: `clusteringMetrics`**Description**: Populated for clustering models.
    - `clusters`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `clusters`**Description**: Information for all clusters.
      - `centroid_id`**Type**: `INT64`**Provider name**: `centroidId`**Description**: Centroid id.
      - `count`**Type**: `INT64`**Provider name**: `count`**Description**: Count of training data rows that were assigned to this cluster.
      - `feature_values`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `featureValues`**Description**: Values of highly variant features for this cluster.
        - `categorical_value`**Type**: `STRUCT`**Provider name**: `categoricalValue`**Description**: The categorical feature value.
          - `category_counts`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `categoryCounts`**Description**: Counts of all categories for the categorical feature. If there are more than ten categories, we return top ten (by count) and return one more CategoryCount with category "*OTHER*" and count as aggregate counts of remaining categories.
            - `category`**Type**: `STRING`**Provider name**: `category`**Description**: The name of category.
            - `count`**Type**: `INT64`**Provider name**: `count`**Description**: The count of training samples matching the category within the cluster.
        - `feature_column`**Type**: `STRING`**Provider name**: `featureColumn`**Description**: The feature column name.
        - `numerical_value`**Type**: `DOUBLE`**Provider name**: `numericalValue`**Description**: The numerical feature value. This is the centroid value for this feature.
    - `davies_bouldin_index`**Type**: `DOUBLE`**Provider name**: `daviesBouldinIndex`**Description**: Davies-Bouldin index.
    - `mean_squared_distance`**Type**: `DOUBLE`**Provider name**: `meanSquaredDistance`**Description**: Mean of squared distances between each sample to its cluster centroid.
  - `dimensionality_reduction_metrics`**Type**: `STRUCT`**Provider name**: `dimensionalityReductionMetrics`**Description**: Evaluation metrics when the model is a dimensionality reduction model, which currently includes PCA.
    - `total_explained_variance_ratio`**Type**: `DOUBLE`**Provider name**: `totalExplainedVarianceRatio`**Description**: Total percentage of variance explained by the selected principal components.
  - `multi_class_classification_metrics`**Type**: `STRUCT`**Provider name**: `multiClassClassificationMetrics`**Description**: Populated for multi-class classification/classifier models.
    - `aggregate_classification_metrics`**Type**: `STRUCT`**Provider name**: `aggregateClassificationMetrics`**Description**: Aggregate classification metrics.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: Accuracy is the fraction of predictions given the correct label. For multiclass this is a micro-averaged metric.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The F1 score is an average of recall and precision. For multiclass this is a macro-averaged metric.
      - `log_loss`**Type**: `DOUBLE`**Provider name**: `logLoss`**Description**: Logarithmic Loss. For multiclass this is a macro-averaged metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: Precision is the fraction of actual positive predictions that had positive actual labels. For multiclass this is a macro-averaged metric treating each class as a binary classifier.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: Recall is the fraction of actual positive labels that were given a positive prediction. For multiclass this is a macro-averaged metric.
      - `roc_auc`**Type**: `DOUBLE`**Provider name**: `rocAuc`**Description**: Area Under a ROC Curve. For multiclass this is a macro-averaged metric.
      - `threshold`**Type**: `DOUBLE`**Provider name**: `threshold`**Description**: Threshold at which the metrics are computed. For binary classification models this is the positive class threshold. For multi-class classfication models this is the confidence threshold.
    - `confusion_matrix_list`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `confusionMatrixList`**Description**: Confusion matrix at different thresholds.
      - `confidence_threshold`**Type**: `DOUBLE`**Provider name**: `confidenceThreshold`**Description**: Confidence threshold used when computing the entries of the confusion matrix.
      - `rows`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `rows`**Description**: One row per actual label.
        - `actual_label`**Type**: `STRING`**Provider name**: `actualLabel`**Description**: The original label of this row.
        - `entries`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `entries`**Description**: Info describing predicted label distribution.
          - `item_count`**Type**: `INT64`**Provider name**: `itemCount`**Description**: Number of items being predicted as this label.
          - `predicted_label`**Type**: `STRING`**Provider name**: `predictedLabel`**Description**: The predicted label. For confidence_threshold > 0, we will also add an entry indicating the number of items under the confidence threshold.
  - `ranking_metrics`**Type**: `STRUCT`**Provider name**: `rankingMetrics`**Description**: Populated for implicit feedback type matrix factorization models.
    - `average_rank`**Type**: `DOUBLE`**Provider name**: `averageRank`**Description**: Determines the goodness of a ranking by computing the percentile rank from the predicted confidence and dividing it by the original rank.
    - `mean_average_precision`**Type**: `DOUBLE`**Provider name**: `meanAveragePrecision`**Description**: Calculates a precision per user for all the items by ranking them and then averages all the precisions across all the users.
    - `mean_squared_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredError`**Description**: Similar to the mean squared error computed in regression and explicit recommendation models except instead of computing the rating directly, the output from evaluate is computed against a preference which is 1 or 0 depending on if the rating exists or not.
    - `normalized_discounted_cumulative_gain`**Type**: `DOUBLE`**Provider name**: `normalizedDiscountedCumulativeGain`**Description**: A metric to determine the goodness of a ranking calculated from the predicted confidence by comparing it to an ideal rank measured by the original ratings.
  - `regression_metrics`**Type**: `STRUCT`**Provider name**: `regressionMetrics`**Description**: Populated for regression models and explicit feedback type matrix factorization models.
    - `mean_absolute_error`**Type**: `DOUBLE`**Provider name**: `meanAbsoluteError`**Description**: Mean absolute error.
    - `mean_squared_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredError`**Description**: Mean squared error.
    - `mean_squared_log_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredLogError`**Description**: Mean squared log error.
    - `median_absolute_error`**Type**: `DOUBLE`**Provider name**: `medianAbsoluteError`**Description**: Median absolute error.
    - `r_squared`**Type**: `DOUBLE`**Provider name**: `rSquared`**Description**: R^2 score. This corresponds to r2_score in ML.EVALUATE.
- `gcp_status`**Type**: `STRING`**Provider name**: `status`**Description**: The status of the trial.**Possible values**:
  - `TRIAL_STATUS_UNSPECIFIED`
  - `NOT_STARTED` - Scheduled but not started.
  - `RUNNING` - Running state.
  - `SUCCEEDED` - The trial succeeded.
  - `FAILED` - The trial failed.
  - `INFEASIBLE` - The trial is infeasible due to the invalid params.
  - `STOPPED_EARLY` - Trial stopped early because it's not promising.
- `hparam_tuning_evaluation_metrics`**Type**: `STRUCT`**Provider name**: `hparamTuningEvaluationMetrics`**Description**: Hyperparameter tuning evaluation metrics of this trial calculated on the eval data. Unlike evaluation_metrics, only the fields corresponding to the hparam_tuning_objectives are set.
  - `arima_forecasting_metrics`**Type**: `STRUCT`**Provider name**: `arimaForecastingMetrics`**Description**: Populated for ARIMA models.
    - `arima_fitting_metrics`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `arimaFittingMetrics`**Description**: Arima model fitting metrics.
      - `aic`**Type**: `DOUBLE`**Provider name**: `aic`**Description**: AIC.
      - `log_likelihood`**Type**: `DOUBLE`**Provider name**: `logLikelihood`**Description**: Log-likelihood.
      - `variance`**Type**: `DOUBLE`**Provider name**: `variance`**Description**: Variance.
    - `arima_single_model_forecasting_metrics`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `arimaSingleModelForecastingMetrics`**Description**: Repeated as there can be many metric sets (one for each model) in auto-arima and the large-scale case.
      - `arima_fitting_metrics`**Type**: `STRUCT`**Provider name**: `arimaFittingMetrics`**Description**: Arima fitting metrics.
        - `aic`**Type**: `DOUBLE`**Provider name**: `aic`**Description**: AIC.
        - `log_likelihood`**Type**: `DOUBLE`**Provider name**: `logLikelihood`**Description**: Log-likelihood.
        - `variance`**Type**: `DOUBLE`**Provider name**: `variance`**Description**: Variance.
      - `has_drift`**Type**: `BOOLEAN`**Provider name**: `hasDrift`**Description**: Is arima model fitted with drift or not. It is always false when d is not 1.
      - `has_holiday_effect`**Type**: `BOOLEAN`**Provider name**: `hasHolidayEffect`**Description**: If true, holiday_effect is a part of time series decomposition result.
      - `has_spikes_and_dips`**Type**: `BOOLEAN`**Provider name**: `hasSpikesAndDips`**Description**: If true, spikes_and_dips is a part of time series decomposition result.
      - `has_step_changes`**Type**: `BOOLEAN`**Provider name**: `hasStepChanges`**Description**: If true, step_changes is a part of time series decomposition result.
      - `non_seasonal_order`**Type**: `STRUCT`**Provider name**: `nonSeasonalOrder`**Description**: Non-seasonal order.
        - `d`**Type**: `INT64`**Provider name**: `d`**Description**: Order of the differencing part.
        - `p`**Type**: `INT64`**Provider name**: `p`**Description**: Order of the autoregressive part.
        - `q`**Type**: `INT64`**Provider name**: `q`**Description**: Order of the moving-average part.
      - `seasonal_periods`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `seasonalPeriods`**Description**: Seasonal periods. Repeated because multiple periods are supported for one time series.
      - `time_series_id`**Type**: `STRING`**Provider name**: `timeSeriesId`**Description**: The time_series_id value for this time series. It will be one of the unique values from the time_series_id_column specified during ARIMA model training. Only present when time_series_id_column training option was used.
      - `time_series_ids`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `timeSeriesIds`**Description**: The tuple of time_series_ids identifying this time series. It will be one of the unique tuples of values present in the time_series_id_columns specified during ARIMA model training. Only present when time_series_id_columns training option was used and the order of values here are same as the order of time_series_id_columns.
    - `has_drift`**Type**: `UNORDERED_LIST_BOOLEAN`**Provider name**: `hasDrift`**Description**: Whether Arima model fitted with drift or not. It is always false when d is not 1.
    - `non_seasonal_order`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `nonSeasonalOrder`**Description**: Non-seasonal order.
      - `d`**Type**: `INT64`**Provider name**: `d`**Description**: Order of the differencing part.
      - `p`**Type**: `INT64`**Provider name**: `p`**Description**: Order of the autoregressive part.
      - `q`**Type**: `INT64`**Provider name**: `q`**Description**: Order of the moving-average part.
    - `seasonal_periods`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `seasonalPeriods`**Description**: Seasonal periods. Repeated because multiple periods are supported for one time series.
    - `time_series_id`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `timeSeriesId`**Description**: Id to differentiate different time series for the large-scale case.
  - `binary_classification_metrics`**Type**: `STRUCT`**Provider name**: `binaryClassificationMetrics`**Description**: Populated for binary classification/classifier models.
    - `aggregate_classification_metrics`**Type**: `STRUCT`**Provider name**: `aggregateClassificationMetrics`**Description**: Aggregate classification metrics.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: Accuracy is the fraction of predictions given the correct label. For multiclass this is a micro-averaged metric.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The F1 score is an average of recall and precision. For multiclass this is a macro-averaged metric.
      - `log_loss`**Type**: `DOUBLE`**Provider name**: `logLoss`**Description**: Logarithmic Loss. For multiclass this is a macro-averaged metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: Precision is the fraction of actual positive predictions that had positive actual labels. For multiclass this is a macro-averaged metric treating each class as a binary classifier.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: Recall is the fraction of actual positive labels that were given a positive prediction. For multiclass this is a macro-averaged metric.
      - `roc_auc`**Type**: `DOUBLE`**Provider name**: `rocAuc`**Description**: Area Under a ROC Curve. For multiclass this is a macro-averaged metric.
      - `threshold`**Type**: `DOUBLE`**Provider name**: `threshold`**Description**: Threshold at which the metrics are computed. For binary classification models this is the positive class threshold. For multi-class classfication models this is the confidence threshold.
    - `binary_confusion_matrix_list`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `binaryConfusionMatrixList`**Description**: Binary confusion matrix at multiple thresholds.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: The fraction of predictions given the correct label.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The equally weighted average of recall and precision.
      - `false_negatives`**Type**: `INT64`**Provider name**: `falseNegatives`**Description**: Number of false samples predicted as false.
      - `false_positives`**Type**: `INT64`**Provider name**: `falsePositives`**Description**: Number of false samples predicted as true.
      - `positive_class_threshold`**Type**: `DOUBLE`**Provider name**: `positiveClassThreshold`**Description**: Threshold value used when computing each of the following metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: The fraction of actual positive predictions that had positive actual labels.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: The fraction of actual positive labels that were given a positive prediction.
      - `true_negatives`**Type**: `INT64`**Provider name**: `trueNegatives`**Description**: Number of true samples predicted as false.
      - `true_positives`**Type**: `INT64`**Provider name**: `truePositives`**Description**: Number of true samples predicted as true.
    - `negative_label`**Type**: `STRING`**Provider name**: `negativeLabel`**Description**: Label representing the negative class.
    - `positive_label`**Type**: `STRING`**Provider name**: `positiveLabel`**Description**: Label representing the positive class.
  - `clustering_metrics`**Type**: `STRUCT`**Provider name**: `clusteringMetrics`**Description**: Populated for clustering models.
    - `clusters`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `clusters`**Description**: Information for all clusters.
      - `centroid_id`**Type**: `INT64`**Provider name**: `centroidId`**Description**: Centroid id.
      - `count`**Type**: `INT64`**Provider name**: `count`**Description**: Count of training data rows that were assigned to this cluster.
      - `feature_values`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `featureValues`**Description**: Values of highly variant features for this cluster.
        - `categorical_value`**Type**: `STRUCT`**Provider name**: `categoricalValue`**Description**: The categorical feature value.
          - `category_counts`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `categoryCounts`**Description**: Counts of all categories for the categorical feature. If there are more than ten categories, we return top ten (by count) and return one more CategoryCount with category "*OTHER*" and count as aggregate counts of remaining categories.
            - `category`**Type**: `STRING`**Provider name**: `category`**Description**: The name of category.
            - `count`**Type**: `INT64`**Provider name**: `count`**Description**: The count of training samples matching the category within the cluster.
        - `feature_column`**Type**: `STRING`**Provider name**: `featureColumn`**Description**: The feature column name.
        - `numerical_value`**Type**: `DOUBLE`**Provider name**: `numericalValue`**Description**: The numerical feature value. This is the centroid value for this feature.
    - `davies_bouldin_index`**Type**: `DOUBLE`**Provider name**: `daviesBouldinIndex`**Description**: Davies-Bouldin index.
    - `mean_squared_distance`**Type**: `DOUBLE`**Provider name**: `meanSquaredDistance`**Description**: Mean of squared distances between each sample to its cluster centroid.
  - `dimensionality_reduction_metrics`**Type**: `STRUCT`**Provider name**: `dimensionalityReductionMetrics`**Description**: Evaluation metrics when the model is a dimensionality reduction model, which currently includes PCA.
    - `total_explained_variance_ratio`**Type**: `DOUBLE`**Provider name**: `totalExplainedVarianceRatio`**Description**: Total percentage of variance explained by the selected principal components.
  - `multi_class_classification_metrics`**Type**: `STRUCT`**Provider name**: `multiClassClassificationMetrics`**Description**: Populated for multi-class classification/classifier models.
    - `aggregate_classification_metrics`**Type**: `STRUCT`**Provider name**: `aggregateClassificationMetrics`**Description**: Aggregate classification metrics.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: Accuracy is the fraction of predictions given the correct label. For multiclass this is a micro-averaged metric.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The F1 score is an average of recall and precision. For multiclass this is a macro-averaged metric.
      - `log_loss`**Type**: `DOUBLE`**Provider name**: `logLoss`**Description**: Logarithmic Loss. For multiclass this is a macro-averaged metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: Precision is the fraction of actual positive predictions that had positive actual labels. For multiclass this is a macro-averaged metric treating each class as a binary classifier.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: Recall is the fraction of actual positive labels that were given a positive prediction. For multiclass this is a macro-averaged metric.
      - `roc_auc`**Type**: `DOUBLE`**Provider name**: `rocAuc`**Description**: Area Under a ROC Curve. For multiclass this is a macro-averaged metric.
      - `threshold`**Type**: `DOUBLE`**Provider name**: `threshold`**Description**: Threshold at which the metrics are computed. For binary classification models this is the positive class threshold. For multi-class classfication models this is the confidence threshold.
    - `confusion_matrix_list`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `confusionMatrixList`**Description**: Confusion matrix at different thresholds.
      - `confidence_threshold`**Type**: `DOUBLE`**Provider name**: `confidenceThreshold`**Description**: Confidence threshold used when computing the entries of the confusion matrix.
      - `rows`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `rows`**Description**: One row per actual label.
        - `actual_label`**Type**: `STRING`**Provider name**: `actualLabel`**Description**: The original label of this row.
        - `entries`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `entries`**Description**: Info describing predicted label distribution.
          - `item_count`**Type**: `INT64`**Provider name**: `itemCount`**Description**: Number of items being predicted as this label.
          - `predicted_label`**Type**: `STRING`**Provider name**: `predictedLabel`**Description**: The predicted label. For confidence_threshold > 0, we will also add an entry indicating the number of items under the confidence threshold.
  - `ranking_metrics`**Type**: `STRUCT`**Provider name**: `rankingMetrics`**Description**: Populated for implicit feedback type matrix factorization models.
    - `average_rank`**Type**: `DOUBLE`**Provider name**: `averageRank`**Description**: Determines the goodness of a ranking by computing the percentile rank from the predicted confidence and dividing it by the original rank.
    - `mean_average_precision`**Type**: `DOUBLE`**Provider name**: `meanAveragePrecision`**Description**: Calculates a precision per user for all the items by ranking them and then averages all the precisions across all the users.
    - `mean_squared_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredError`**Description**: Similar to the mean squared error computed in regression and explicit recommendation models except instead of computing the rating directly, the output from evaluate is computed against a preference which is 1 or 0 depending on if the rating exists or not.
    - `normalized_discounted_cumulative_gain`**Type**: `DOUBLE`**Provider name**: `normalizedDiscountedCumulativeGain`**Description**: A metric to determine the goodness of a ranking calculated from the predicted confidence by comparing it to an ideal rank measured by the original ratings.
  - `regression_metrics`**Type**: `STRUCT`**Provider name**: `regressionMetrics`**Description**: Populated for regression models and explicit feedback type matrix factorization models.
    - `mean_absolute_error`**Type**: `DOUBLE`**Provider name**: `meanAbsoluteError`**Description**: Mean absolute error.
    - `mean_squared_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredError`**Description**: Mean squared error.
    - `mean_squared_log_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredLogError`**Description**: Mean squared log error.
    - `median_absolute_error`**Type**: `DOUBLE`**Provider name**: `medianAbsoluteError`**Description**: Median absolute error.
    - `r_squared`**Type**: `DOUBLE`**Provider name**: `rSquared`**Description**: R^2 score. This corresponds to r2_score in ML.EVALUATE.
- `hparams`**Type**: `STRUCT`**Provider name**: `hparams`**Description**: The hyperprameters selected for this trial.
  - `adjust_step_changes`**Type**: `BOOLEAN`**Provider name**: `adjustStepChanges`**Description**: If true, detect step changes and make data adjustment in the input time series.
  - `approx_global_feature_contrib`**Type**: `BOOLEAN`**Provider name**: `approxGlobalFeatureContrib`**Description**: Whether to use approximate feature contribution method in XGBoost model explanation for global explain.
  - `auto_arima`**Type**: `BOOLEAN`**Provider name**: `autoArima`**Description**: Whether to enable auto ARIMA or not.
  - `auto_arima_max_order`**Type**: `INT64`**Provider name**: `autoArimaMaxOrder`**Description**: The max value of the sum of non-seasonal p and q.
  - `auto_arima_min_order`**Type**: `INT64`**Provider name**: `autoArimaMinOrder`**Description**: The min value of the sum of non-seasonal p and q.
  - `batch_size`**Type**: `INT64`**Provider name**: `batchSize`**Description**: Batch size for dnn models.
  - `booster_type`**Type**: `STRING`**Provider name**: `boosterType`**Description**: Booster type for boosted tree models.**Possible values**:
    - `BOOSTER_TYPE_UNSPECIFIED` - Unspecified booster type.
    - `GBTREE` - Gbtree booster.
    - `DART` - Dart booster.
  - `calculate_p_values`**Type**: `BOOLEAN`**Provider name**: `calculatePValues`**Description**: Whether or not p-value test should be computed for this model. Only available for linear and logistic regression models.
  - `clean_spikes_and_dips`**Type**: `BOOLEAN`**Provider name**: `cleanSpikesAndDips`**Description**: If true, clean spikes and dips in the input time series.
  - `color_space`**Type**: `STRING`**Provider name**: `colorSpace`**Description**: Enums for color space, used for processing images in Object Table. See more details at [https://www.tensorflow.org/io/tutorials/colorspace](https://www.tensorflow.org/io/tutorials/colorspace).**Possible values**:
    - `COLOR_SPACE_UNSPECIFIED` - Unspecified color space
    - `RGB` - RGB
    - `HSV` - HSV
    - `YIQ` - YIQ
    - `YUV` - YUV
    - `GRAYSCALE` - GRAYSCALE
  - `colsample_bylevel`**Type**: `DOUBLE`**Provider name**: `colsampleBylevel`**Description**: Subsample ratio of columns for each level for boosted tree models.
  - `colsample_bynode`**Type**: `DOUBLE`**Provider name**: `colsampleBynode`**Description**: Subsample ratio of columns for each node(split) for boosted tree models.
  - `colsample_bytree`**Type**: `DOUBLE`**Provider name**: `colsampleBytree`**Description**: Subsample ratio of columns when constructing each tree for boosted tree models.
  - `dart_normalize_type`**Type**: `STRING`**Provider name**: `dartNormalizeType`**Description**: Type of normalization algorithm for boosted tree models using dart booster.**Possible values**:
    - `DART_NORMALIZE_TYPE_UNSPECIFIED` - Unspecified dart normalize type.
    - `TREE` - New trees have the same weight of each of dropped trees.
    - `FOREST` - New trees have the same weight of sum of dropped trees.
  - `data_frequency`**Type**: `STRING`**Provider name**: `dataFrequency`**Description**: The data frequency of a time series.**Possible values**:
    - `DATA_FREQUENCY_UNSPECIFIED`
    - `AUTO_FREQUENCY` - Automatically inferred from timestamps.
    - `YEARLY` - Yearly data.
    - `QUARTERLY` - Quarterly data.
    - `MONTHLY` - Monthly data.
    - `WEEKLY` - Weekly data.
    - `DAILY` - Daily data.
    - `HOURLY` - Hourly data.
    - `PER_MINUTE` - Per-minute data.
  - `data_split_column`**Type**: `STRING`**Provider name**: `dataSplitColumn`**Description**: The column to split data with. This column won't be used as a feature. 1. When data_split_method is CUSTOM, the corresponding column should be boolean. The rows with true value tag are eval data, and the false are training data. 2. When data_split_method is SEQ, the first DATA_SPLIT_EVAL_FRACTION rows (from smallest to largest) in the corresponding column are used as training data, and the rest are eval data. It respects the order in Orderable data types: [https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#data-type-properties](https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#data-type-properties)
  - `data_split_eval_fraction`**Type**: `DOUBLE`**Provider name**: `dataSplitEvalFraction`**Description**: The fraction of evaluation data over the whole input data. The rest of data will be used as training data. The format should be double. Accurate to two decimal places. Default value is 0.2.
  - `data_split_method`**Type**: `STRING`**Provider name**: `dataSplitMethod`**Description**: The data split type for training and evaluation, e.g. RANDOM.**Possible values**:
    - `DATA_SPLIT_METHOD_UNSPECIFIED`
    - `RANDOM` - Splits data randomly.
    - `CUSTOM` - Splits data with the user provided tags.
    - `SEQUENTIAL` - Splits data sequentially.
    - `NO_SPLIT` - Data split will be skipped.
    - `AUTO_SPLIT` - Splits data automatically: Uses NO_SPLIT if the data size is small. Otherwise uses RANDOM.
  - `decompose_time_series`**Type**: `BOOLEAN`**Provider name**: `decomposeTimeSeries`**Description**: If true, perform decompose time series and save the results.
  - `distance_type`**Type**: `STRING`**Provider name**: `distanceType`**Description**: Distance type for clustering models.**Possible values**:
    - `DISTANCE_TYPE_UNSPECIFIED`
    - `EUCLIDEAN` - Eculidean distance.
    - `COSINE` - Cosine distance.
  - `dropout`**Type**: `DOUBLE`**Provider name**: `dropout`**Description**: Dropout probability for dnn models.
  - `early_stop`**Type**: `BOOLEAN`**Provider name**: `earlyStop`**Description**: Whether to stop early when the loss doesn't improve significantly any more (compared to min_relative_progress). Used only for iterative training algorithms.
  - `enable_global_explain`**Type**: `BOOLEAN`**Provider name**: `enableGlobalExplain`**Description**: If true, enable global explanation during training.
  - `feedback_type`**Type**: `STRING`**Provider name**: `feedbackType`**Description**: Feedback type that specifies which algorithm to run for matrix factorization.**Possible values**:
    - `FEEDBACK_TYPE_UNSPECIFIED`
    - `IMPLICIT` - Use weighted-als for implicit feedback problems.
    - `EXPLICIT` - Use nonweighted-als for explicit feedback problems.
  - `hidden_units`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `hiddenUnits`**Description**: Hidden units for dnn models.
  - `holiday_region`**Type**: `STRING`**Provider name**: `holidayRegion`**Description**: The geographical region based on which the holidays are considered in time series modeling. If a valid value is specified, then holiday effects modeling is enabled.**Possible values**:
    - `HOLIDAY_REGION_UNSPECIFIED` - Holiday region unspecified.
    - `GLOBAL` - Global.
    - `NA` - North America.
    - `JAPAC` - Japan and Asia Pacific: Korea, Greater China, India, Australia, and New Zealand.
    - `EMEA` - Europe, the Middle East and Africa.
    - `LAC` - Latin America and the Caribbean.
    - `AE` - United Arab Emirates
    - `AR` - Argentina
    - `AT` - Austria
    - `AU` - Australia
    - `BE` - Belgium
    - `BR` - Brazil
    - `CA` - Canada
    - `CH` - Switzerland
    - `CL` - Chile
    - `CN` - China
    - `CO` - Colombia
    - `CS` - Czechoslovakia
    - `CZ` - Czech Republic
    - `DE` - Germany
    - `DK` - Denmark
    - `DZ` - Algeria
    - `EC` - Ecuador
    - `EE` - Estonia
    - `EG` - Egypt
    - `ES` - Spain
    - `FI` - Finland
    - `FR` - France
    - `GB` - Great Britain (United Kingdom)
    - `GR` - Greece
    - `HK` - Hong Kong
    - `HU` - Hungary
    - `ID` - Indonesia
    - `IE` - Ireland
    - `IL` - Israel
    - `IN` - India
    - `IR` - Iran
    - `IT` - Italy
    - `JP` - Japan
    - `KR` - Korea (South)
    - `LV` - Latvia
    - `MA` - Morocco
    - `MX` - Mexico
    - `MY` - Malaysia
    - `NG` - Nigeria
    - `NL` - Netherlands
    - `NO` - Norway
    - `NZ` - New Zealand
    - `PE` - Peru
    - `PH` - Philippines
    - `PK` - Pakistan
    - `PL` - Poland
    - `PT` - Portugal
    - `RO` - Romania
    - `RS` - Serbia
    - `RU` - Russian Federation
    - `SA` - Saudi Arabia
    - `SE` - Sweden
    - `SG` - Singapore
    - `SI` - Slovenia
    - `SK` - Slovakia
    - `TH` - Thailand
    - `TR` - Turkey
    - `TW` - Taiwan
    - `UA` - Ukraine
    - `US` - United States
    - `VE` - Venezuela
    - `VN` - Viet Nam
    - `ZA` - South Africa
  - `horizon`**Type**: `INT64`**Provider name**: `horizon`**Description**: The number of periods ahead that need to be forecasted.
  - `hparam_tuning_objectives`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `hparamTuningObjectives`**Description**: The target evaluation metrics to optimize the hyperparameters for.
  - `include_drift`**Type**: `BOOLEAN`**Provider name**: `includeDrift`**Description**: Include drift when fitting an ARIMA model.
  - `initial_learn_rate`**Type**: `DOUBLE`**Provider name**: `initialLearnRate`**Description**: Specifies the initial learning rate for the line search learn rate strategy.
  - `input_label_columns`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `inputLabelColumns`**Description**: Name of input label columns in training data.
  - `instance_weight_column`**Type**: `STRING`**Provider name**: `instanceWeightColumn`**Description**: Name of the instance weight column for training data. This column isn't be used as a feature.
  - `integrated_gradients_num_steps`**Type**: `INT64`**Provider name**: `integratedGradientsNumSteps`**Description**: Number of integral steps for the integrated gradients explain method.
  - `item_column`**Type**: `STRING`**Provider name**: `itemColumn`**Description**: Item column specified for matrix factorization models.
  - `kmeans_initialization_column`**Type**: `STRING`**Provider name**: `kmeansInitializationColumn`**Description**: The column used to provide the initial centroids for kmeans algorithm when kmeans_initialization_method is CUSTOM.
  - `kmeans_initialization_method`**Type**: `STRING`**Provider name**: `kmeansInitializationMethod`**Description**: The method used to initialize the centroids for kmeans algorithm.**Possible values**:
    - `KMEANS_INITIALIZATION_METHOD_UNSPECIFIED` - Unspecified initialization method.
    - `RANDOM` - Initializes the centroids randomly.
    - `CUSTOM` - Initializes the centroids using data specified in kmeans_initialization_column.
    - `KMEANS_PLUS_PLUS` - Initializes with kmeans++.
  - `l1_regularization`**Type**: `DOUBLE`**Provider name**: `l1Regularization`**Description**: L1 regularization coefficient.
  - `l2_regularization`**Type**: `DOUBLE`**Provider name**: `l2Regularization`**Description**: L2 regularization coefficient.
  - `learn_rate`**Type**: `DOUBLE`**Provider name**: `learnRate`**Description**: Learning rate in training. Used only for iterative training algorithms.
  - `learn_rate_strategy`**Type**: `STRING`**Provider name**: `learnRateStrategy`**Description**: The strategy to determine learn rate for the current iteration.**Possible values**:
    - `LEARN_RATE_STRATEGY_UNSPECIFIED`
    - `LINE_SEARCH` - Use line search to determine learning rate.
    - `CONSTANT` - Use a constant learning rate.
  - `loss_type`**Type**: `STRING`**Provider name**: `lossType`**Description**: Type of loss function used during training run.**Possible values**:
    - `LOSS_TYPE_UNSPECIFIED`
    - `MEAN_SQUARED_LOSS` - Mean squared loss, used for linear regression.
    - `MEAN_LOG_LOSS` - Mean log loss, used for logistic regression.
  - `max_iterations`**Type**: `INT64`**Provider name**: `maxIterations`**Description**: The maximum number of iterations in training. Used only for iterative training algorithms.
  - `max_parallel_trials`**Type**: `INT64`**Provider name**: `maxParallelTrials`**Description**: Maximum number of trials to run in parallel.
  - `max_time_series_length`**Type**: `INT64`**Provider name**: `maxTimeSeriesLength`**Description**: Get truncated length by last n points in time series. Use separately from time_series_length_fraction and min_time_series_length.
  - `max_tree_depth`**Type**: `INT64`**Provider name**: `maxTreeDepth`**Description**: Maximum depth of a tree for boosted tree models.
  - `min_relative_progress`**Type**: `DOUBLE`**Provider name**: `minRelativeProgress`**Description**: When early_stop is true, stops training when accuracy improvement is less than 'min_relative_progress'. Used only for iterative training algorithms.
  - `min_split_loss`**Type**: `DOUBLE`**Provider name**: `minSplitLoss`**Description**: Minimum split loss for boosted tree models.
  - `min_time_series_length`**Type**: `INT64`**Provider name**: `minTimeSeriesLength`**Description**: Set fast trend ARIMA_PLUS model minimum training length. Use in pair with time_series_length_fraction.
  - `min_tree_child_weight`**Type**: `INT64`**Provider name**: `minTreeChildWeight`**Description**: Minimum sum of instance weight needed in a child for boosted tree models.
  - `model_uri`**Type**: `STRING`**Provider name**: `modelUri`**Description**: Google Cloud Storage URI from which the model was imported. Only applicable for imported models.
  - `non_seasonal_order`**Type**: `STRUCT`**Provider name**: `nonSeasonalOrder`**Description**: A specification of the non-seasonal part of the ARIMA model: the three components (p, d, q) are the AR order, the degree of differencing, and the MA order.
    - `d`**Type**: `INT64`**Provider name**: `d`**Description**: Order of the differencing part.
    - `p`**Type**: `INT64`**Provider name**: `p`**Description**: Order of the autoregressive part.
    - `q`**Type**: `INT64`**Provider name**: `q`**Description**: Order of the moving-average part.
  - `num_clusters`**Type**: `INT64`**Provider name**: `numClusters`**Description**: Number of clusters for clustering models.
  - `num_factors`**Type**: `INT64`**Provider name**: `numFactors`**Description**: Num factors specified for matrix factorization models.
  - `num_parallel_tree`**Type**: `INT64`**Provider name**: `numParallelTree`**Description**: Number of parallel trees constructed during each iteration for boosted tree models.
  - `num_trials`**Type**: `INT64`**Provider name**: `numTrials`**Description**: Number of trials to run this hyperparameter tuning job.
  - `optimization_strategy`**Type**: `STRING`**Provider name**: `optimizationStrategy`**Description**: Optimization strategy for training linear regression models.**Possible values**:
    - `OPTIMIZATION_STRATEGY_UNSPECIFIED`
    - `BATCH_GRADIENT_DESCENT` - Uses an iterative batch gradient descent algorithm.
    - `NORMAL_EQUATION` - Uses a normal equation to solve linear regression problem.
  - `sampled_shapley_num_paths`**Type**: `INT64`**Provider name**: `sampledShapleyNumPaths`**Description**: Number of paths for the sampled Shapley explain method.
  - `subsample`**Type**: `DOUBLE`**Provider name**: `subsample`**Description**: Subsample fraction of the training data to grow tree to prevent overfitting for boosted tree models.
  - `tf_version`**Type**: `STRING`**Provider name**: `tfVersion`**Description**: Based on the selected TF version, the corresponding docker image is used to train external models.
  - `time_series_data_column`**Type**: `STRING`**Provider name**: `timeSeriesDataColumn`**Description**: Column to be designated as time series data for ARIMA model.
  - `time_series_id_column`**Type**: `STRING`**Provider name**: `timeSeriesIdColumn`**Description**: The time series id column that was used during ARIMA model training.
  - `time_series_id_columns`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `timeSeriesIdColumns`**Description**: The time series id columns that were used during ARIMA model training.
  - `time_series_length_fraction`**Type**: `DOUBLE`**Provider name**: `timeSeriesLengthFraction`**Description**: Get truncated length by fraction in time series.
  - `time_series_timestamp_column`**Type**: `STRING`**Provider name**: `timeSeriesTimestampColumn`**Description**: Column to be designated as time series timestamp for ARIMA model.
  - `tree_method`**Type**: `STRING`**Provider name**: `treeMethod`**Description**: Tree construction algorithm for boosted tree models.**Possible values**:
    - `TREE_METHOD_UNSPECIFIED` - Unspecified tree method.
    - `AUTO` - Use heuristic to choose the fastest method.
    - `EXACT` - Exact greedy algorithm.
    - `APPROX` - Approximate greedy algorithm using quantile sketch and gradient histogram.
    - `HIST` - Fast histogram optimized approximate greedy algorithm.
  - `trend_smoothing_window_size`**Type**: `INT64`**Provider name**: `trendSmoothingWindowSize`**Description**: The smoothing window size for the trend component of the time series.
  - `user_column`**Type**: `STRING`**Provider name**: `userColumn`**Description**: User column specified for matrix factorization models.
  - `wals_alpha`**Type**: `DOUBLE`**Provider name**: `walsAlpha`**Description**: Hyperparameter for matrix factoration when implicit feedback type is specified.
  - `warm_start`**Type**: `BOOLEAN`**Provider name**: `warmStart`**Description**: Whether to train a model from the last checkpoint.
  - `xgboost_version`**Type**: `STRING`**Provider name**: `xgboostVersion`**Description**: User-selected XGBoost versions for training of XGBoost models.
- `start_time_ms`**Type**: `INT64`**Provider name**: `startTimeMs`**Description**: Starting time of the trial.
- `training_loss`**Type**: `DOUBLE`**Provider name**: `trainingLoss`**Description**: Loss computed on the training data at the end of trial.
- `trial_id`**Type**: `INT64`**Provider name**: `trialId`**Description**: 1-based index of the trial.

## `label_columns`{% #label_columns %}

**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `labelColumns`**Description**: Output only. Label columns that were used to train this model. The output of the model will have a "predicted_" prefix to these columns.

- `name`**Type**: `STRING`**Provider name**: `name`**Description**: Optional. The name of this field. Can be absent for struct fields.
- `type`**Type**: `STRUCT`**Provider name**: `type`**Description**: Optional. The type of this parameter. Absent if not explicitly specified (e.g., CREATE FUNCTION statement can omit the return type; in this case the output parameter does not have this "type" field).
  - `struct_type`**Type**: `STRUCT`**Provider name**: `structType`**Description**: The fields of this struct, in order, if type_kind = "STRUCT".

  - `type_kind`**Type**: `STRING`**Provider name**: `typeKind`**Description**: Required. The top level type of this field. Can be any GoogleSQL data type (e.g., "INT64", "DATE", "ARRAY").**Possible values**:

    - `TYPE_KIND_UNSPECIFIED` - Invalid type.
    - `INT64` - Encoded as a string in decimal format.
    - `BOOL` - Encoded as a boolean 'false' or 'true'.
    - `FLOAT64` - Encoded as a number, or string 'NaN
    - `STRING` - Infinity' or '-Infinity'.
    - `BYTES` - Encoded as a string value.
    - `TIMESTAMP` - Encoded as a base64 string per RFC 4648, section 4.
    - `DATE` - Encoded as an RFC 3339 timestamp with mandatory 'Z' time zone string: 1985-04-12T23:20:50.52Z
    - `TIME` - Encoded as RFC 3339 full-date format string: 1985-04-12
    - `DATETIME` - Encoded as RFC 3339 partial-time format string: 23:20:50.52
    - `INTERVAL` - Encoded as RFC 3339 full-date 'T' partial-time: 1985-04-12T23:20:50.52
    - `GEOGRAPHY` - Encoded as fully qualified 3 part: 0-5 15 2:30:45.6
    - `NUMERIC` - Encoded as WKT
    - `BIGNUMERIC` - Encoded as a decimal string.
    - `JSON` - Encoded as a decimal string.
    - `ARRAY` - Encoded as a string.
    - `STRUCT` - Encoded as a list with types matching Type.array_type.

## `labels`{% #labels %}

**Type**: `UNORDERED_LIST_STRING`

## `last_modified_time`{% #last_modified_time %}

**Type**: `INT64`**Provider name**: `lastModifiedTime`**Description**: Output only. The time when this model was last modified, in millisecs since the epoch.

## `location`{% #location %}

**Type**: `STRING`**Provider name**: `location`**Description**: Output only. The geographic location where the model resides. This value is inherited from the dataset.

## `model_reference`{% #model_reference %}

**Type**: `STRUCT`**Provider name**: `modelReference`**Description**: Required. Unique identifier for this model.

- `dataset_id`**Type**: `STRING`**Provider name**: `datasetId`**Description**: Required. The ID of the dataset containing this model.
- `model_id`**Type**: `STRING`**Provider name**: `modelId`**Description**: Required. The ID of the model. The ID must contain only letters (a-z, A-Z), numbers (0-9), or underscores (_). The maximum length is 1,024 characters.
- `project_id`**Type**: `STRING`**Provider name**: `projectId`**Description**: Required. The ID of the project containing this model.

## `model_type`{% #model_type %}

**Type**: `STRING`**Provider name**: `modelType`**Description**: Output only. Type of the model resource.**Possible values**:

- `MODEL_TYPE_UNSPECIFIED`
- `LINEAR_REGRESSION` - Linear regression model.
- `LOGISTIC_REGRESSION` - Logistic regression based classification model.
- `KMEANS` - K-means clustering model.
- `MATRIX_FACTORIZATION` - Matrix factorization model.
- `DNN_CLASSIFIER` - DNN classifier model.
- `TENSORFLOW` - An imported TensorFlow model.
- `DNN_REGRESSOR` - DNN regressor model.
- `XGBOOST` - An imported XGBoost model.
- `BOOSTED_TREE_REGRESSOR` - Boosted tree regressor model.
- `BOOSTED_TREE_CLASSIFIER` - Boosted tree classifier model.
- `ARIMA` - ARIMA model.
- `AUTOML_REGRESSOR` - AutoML Tables regression model.
- `AUTOML_CLASSIFIER` - AutoML Tables classification model.
- `PCA` - Prinpical Component Analysis model.
- `DNN_LINEAR_COMBINED_CLASSIFIER` - Wide-and-deep classifier model.
- `DNN_LINEAR_COMBINED_REGRESSOR` - Wide-and-deep regressor model.
- `AUTOENCODER` - Autoencoder model.
- `ARIMA_PLUS` - New name for the ARIMA model.
- `ARIMA_PLUS_XREG` - ARIMA with external regressors.
- `RANDOM_FOREST_REGRESSOR` - Random forest regressor model.
- `RANDOM_FOREST_CLASSIFIER` - Random forest classifier model.
- `TENSORFLOW_LITE` - An imported TensorFlow Lite model.
- `ONNX` - An imported ONNX model.

## `optimal_trial_ids`{% #optimal_trial_ids %}

**Type**: `UNORDERED_LIST_INT64`**Provider name**: `optimalTrialIds`**Description**: Output only. For single-objective [hyperparameter tuning](https://docs.datadoghq.com/bigquery-ml/docs/reference/standard-sql/bigqueryml-syntax-hp-tuning-overview) models, it only contains the best trial. For multi-objective [hyperparameter tuning](https://docs.datadoghq.com/bigquery-ml/docs/reference/standard-sql/bigqueryml-syntax-hp-tuning-overview) models, it contains all Pareto optimal trials sorted by trial_id.

## `organization_id`{% #organization_id %}

**Type**: `STRING`

## `parent`{% #parent %}

**Type**: `STRING`

## `project_id`{% #project_id %}

**Type**: `STRING`

## `project_number`{% #project_number %}

**Type**: `STRING`

## `region_id`{% #region_id %}

**Type**: `STRING`

## `remote_model_info`{% #remote_model_info %}

**Type**: `STRUCT`**Provider name**: `remoteModelInfo`**Description**: Output only. Remote model info

- `connection`**Type**: `STRING`**Provider name**: `connection`**Description**: Output only. Fully qualified name of the user-provided connection object of the remote model. Format: `"projects/{project_id}/locations/{location_id}/connections/{connection_id}"`
- `endpoint`**Type**: `STRING`**Provider name**: `endpoint`**Description**: Output only. The endpoint for remote model.
- `max_batching_rows`**Type**: `INT64`**Provider name**: `maxBatchingRows`**Description**: Output only. Max number of rows in each batch sent to the remote service. If unset, the number of rows in each batch is set dynamically.
- `remote_service_type`**Type**: `STRING`**Provider name**: `remoteServiceType`**Description**: Output only. The remote service type for remote model.**Possible values**:
  - `REMOTE_SERVICE_TYPE_UNSPECIFIED` - Unspecified remote service type.
  - `CLOUD_AI_TRANSLATE_V3` - V3 Cloud AI Translation API. See more details at [Cloud Translation API] ([https://cloud.google.com/translate/docs/reference/rest)](https://cloud.google.com/translate/docs/reference/rest%29).
  - `CLOUD_AI_VISION_V1` - V1 Cloud AI Vision API See more details at [Cloud Vision API] ([https://cloud.google.com/vision/docs/reference/rest)](https://cloud.google.com/vision/docs/reference/rest%29).
  - `CLOUD_AI_NATURAL_LANGUAGE_V1` - V1 Cloud AI Natural Language API. See more details at [REST Resource: documents](https://cloud.google.com/natural-language/docs/reference/rest/v1/documents).

## `resource_name`{% #resource_name %}

**Type**: `STRING`

## `tags`{% #tags %}

**Type**: `UNORDERED_LIST_STRING`

## `training_runs`{% #training_runs %}

**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `trainingRuns`**Description**: Information for all training runs in increasing order of start_time.

- `class_level_global_explanations`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `classLevelGlobalExplanations`**Description**: Output only. Global explanation contains the explanation of top features on the class level. Applies to classification models only.
  - `class_label`**Type**: `STRING`**Provider name**: `classLabel`**Description**: Class label for this set of global explanations. Will be empty/null for binary logistic and linear regression models. Sorted alphabetically in descending order.
  - `explanations`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `explanations`**Description**: A list of the top global explanations. Sorted by absolute value of attribution in descending order.
    - `attribution`**Type**: `DOUBLE`**Provider name**: `attribution`**Description**: Attribution of feature.
    - `feature_name`**Type**: `STRING`**Provider name**: `featureName`**Description**: The full feature name. For non-numerical features, will be formatted like `.`. Overall size of feature name will always be truncated to first 120 characters.
- `data_split_result`**Type**: `STRUCT`**Provider name**: `dataSplitResult`**Description**: Output only. Data split result of the training run. Only set when the input data is actually split.
  - `evaluation_table`**Type**: `STRUCT`**Provider name**: `evaluationTable`**Description**: Table reference of the evaluation data after split.
    - `dataset_id`**Type**: `STRING`**Provider name**: `datasetId`**Description**: [Required] The ID of the dataset containing this table.
    - `project_id`**Type**: `STRING`**Provider name**: `projectId`**Description**: [Required] The ID of the project containing this table.
    - `table_id`**Type**: `STRING`**Provider name**: `tableId`**Description**: [Required] The ID of the table. The ID must contain only letters (a-z, A-Z), numbers (0-9), or underscores (_). The maximum length is 1,024 characters.
  - `test_table`**Type**: `STRUCT`**Provider name**: `testTable`**Description**: Table reference of the test data after split.
    - `dataset_id`**Type**: `STRING`**Provider name**: `datasetId`**Description**: [Required] The ID of the dataset containing this table.
    - `project_id`**Type**: `STRING`**Provider name**: `projectId`**Description**: [Required] The ID of the project containing this table.
    - `table_id`**Type**: `STRING`**Provider name**: `tableId`**Description**: [Required] The ID of the table. The ID must contain only letters (a-z, A-Z), numbers (0-9), or underscores (_). The maximum length is 1,024 characters.
  - `training_table`**Type**: `STRUCT`**Provider name**: `trainingTable`**Description**: Table reference of the training data after split.
    - `dataset_id`**Type**: `STRING`**Provider name**: `datasetId`**Description**: [Required] The ID of the dataset containing this table.
    - `project_id`**Type**: `STRING`**Provider name**: `projectId`**Description**: [Required] The ID of the project containing this table.
    - `table_id`**Type**: `STRING`**Provider name**: `tableId`**Description**: [Required] The ID of the table. The ID must contain only letters (a-z, A-Z), numbers (0-9), or underscores (_). The maximum length is 1,024 characters.
- `evaluation_metrics`**Type**: `STRUCT`**Provider name**: `evaluationMetrics`**Description**: Output only. The evaluation metrics over training/eval data that were computed at the end of training.
  - `arima_forecasting_metrics`**Type**: `STRUCT`**Provider name**: `arimaForecastingMetrics`**Description**: Populated for ARIMA models.
    - `arima_fitting_metrics`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `arimaFittingMetrics`**Description**: Arima model fitting metrics.
      - `aic`**Type**: `DOUBLE`**Provider name**: `aic`**Description**: AIC.
      - `log_likelihood`**Type**: `DOUBLE`**Provider name**: `logLikelihood`**Description**: Log-likelihood.
      - `variance`**Type**: `DOUBLE`**Provider name**: `variance`**Description**: Variance.
    - `arima_single_model_forecasting_metrics`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `arimaSingleModelForecastingMetrics`**Description**: Repeated as there can be many metric sets (one for each model) in auto-arima and the large-scale case.
      - `arima_fitting_metrics`**Type**: `STRUCT`**Provider name**: `arimaFittingMetrics`**Description**: Arima fitting metrics.
        - `aic`**Type**: `DOUBLE`**Provider name**: `aic`**Description**: AIC.
        - `log_likelihood`**Type**: `DOUBLE`**Provider name**: `logLikelihood`**Description**: Log-likelihood.
        - `variance`**Type**: `DOUBLE`**Provider name**: `variance`**Description**: Variance.
      - `has_drift`**Type**: `BOOLEAN`**Provider name**: `hasDrift`**Description**: Is arima model fitted with drift or not. It is always false when d is not 1.
      - `has_holiday_effect`**Type**: `BOOLEAN`**Provider name**: `hasHolidayEffect`**Description**: If true, holiday_effect is a part of time series decomposition result.
      - `has_spikes_and_dips`**Type**: `BOOLEAN`**Provider name**: `hasSpikesAndDips`**Description**: If true, spikes_and_dips is a part of time series decomposition result.
      - `has_step_changes`**Type**: `BOOLEAN`**Provider name**: `hasStepChanges`**Description**: If true, step_changes is a part of time series decomposition result.
      - `non_seasonal_order`**Type**: `STRUCT`**Provider name**: `nonSeasonalOrder`**Description**: Non-seasonal order.
        - `d`**Type**: `INT64`**Provider name**: `d`**Description**: Order of the differencing part.
        - `p`**Type**: `INT64`**Provider name**: `p`**Description**: Order of the autoregressive part.
        - `q`**Type**: `INT64`**Provider name**: `q`**Description**: Order of the moving-average part.
      - `seasonal_periods`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `seasonalPeriods`**Description**: Seasonal periods. Repeated because multiple periods are supported for one time series.
      - `time_series_id`**Type**: `STRING`**Provider name**: `timeSeriesId`**Description**: The time_series_id value for this time series. It will be one of the unique values from the time_series_id_column specified during ARIMA model training. Only present when time_series_id_column training option was used.
      - `time_series_ids`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `timeSeriesIds`**Description**: The tuple of time_series_ids identifying this time series. It will be one of the unique tuples of values present in the time_series_id_columns specified during ARIMA model training. Only present when time_series_id_columns training option was used and the order of values here are same as the order of time_series_id_columns.
    - `has_drift`**Type**: `UNORDERED_LIST_BOOLEAN`**Provider name**: `hasDrift`**Description**: Whether Arima model fitted with drift or not. It is always false when d is not 1.
    - `non_seasonal_order`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `nonSeasonalOrder`**Description**: Non-seasonal order.
      - `d`**Type**: `INT64`**Provider name**: `d`**Description**: Order of the differencing part.
      - `p`**Type**: `INT64`**Provider name**: `p`**Description**: Order of the autoregressive part.
      - `q`**Type**: `INT64`**Provider name**: `q`**Description**: Order of the moving-average part.
    - `seasonal_periods`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `seasonalPeriods`**Description**: Seasonal periods. Repeated because multiple periods are supported for one time series.
    - `time_series_id`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `timeSeriesId`**Description**: Id to differentiate different time series for the large-scale case.
  - `binary_classification_metrics`**Type**: `STRUCT`**Provider name**: `binaryClassificationMetrics`**Description**: Populated for binary classification/classifier models.
    - `aggregate_classification_metrics`**Type**: `STRUCT`**Provider name**: `aggregateClassificationMetrics`**Description**: Aggregate classification metrics.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: Accuracy is the fraction of predictions given the correct label. For multiclass this is a micro-averaged metric.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The F1 score is an average of recall and precision. For multiclass this is a macro-averaged metric.
      - `log_loss`**Type**: `DOUBLE`**Provider name**: `logLoss`**Description**: Logarithmic Loss. For multiclass this is a macro-averaged metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: Precision is the fraction of actual positive predictions that had positive actual labels. For multiclass this is a macro-averaged metric treating each class as a binary classifier.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: Recall is the fraction of actual positive labels that were given a positive prediction. For multiclass this is a macro-averaged metric.
      - `roc_auc`**Type**: `DOUBLE`**Provider name**: `rocAuc`**Description**: Area Under a ROC Curve. For multiclass this is a macro-averaged metric.
      - `threshold`**Type**: `DOUBLE`**Provider name**: `threshold`**Description**: Threshold at which the metrics are computed. For binary classification models this is the positive class threshold. For multi-class classfication models this is the confidence threshold.
    - `binary_confusion_matrix_list`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `binaryConfusionMatrixList`**Description**: Binary confusion matrix at multiple thresholds.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: The fraction of predictions given the correct label.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The equally weighted average of recall and precision.
      - `false_negatives`**Type**: `INT64`**Provider name**: `falseNegatives`**Description**: Number of false samples predicted as false.
      - `false_positives`**Type**: `INT64`**Provider name**: `falsePositives`**Description**: Number of false samples predicted as true.
      - `positive_class_threshold`**Type**: `DOUBLE`**Provider name**: `positiveClassThreshold`**Description**: Threshold value used when computing each of the following metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: The fraction of actual positive predictions that had positive actual labels.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: The fraction of actual positive labels that were given a positive prediction.
      - `true_negatives`**Type**: `INT64`**Provider name**: `trueNegatives`**Description**: Number of true samples predicted as false.
      - `true_positives`**Type**: `INT64`**Provider name**: `truePositives`**Description**: Number of true samples predicted as true.
    - `negative_label`**Type**: `STRING`**Provider name**: `negativeLabel`**Description**: Label representing the negative class.
    - `positive_label`**Type**: `STRING`**Provider name**: `positiveLabel`**Description**: Label representing the positive class.
  - `clustering_metrics`**Type**: `STRUCT`**Provider name**: `clusteringMetrics`**Description**: Populated for clustering models.
    - `clusters`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `clusters`**Description**: Information for all clusters.
      - `centroid_id`**Type**: `INT64`**Provider name**: `centroidId`**Description**: Centroid id.
      - `count`**Type**: `INT64`**Provider name**: `count`**Description**: Count of training data rows that were assigned to this cluster.
      - `feature_values`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `featureValues`**Description**: Values of highly variant features for this cluster.
        - `categorical_value`**Type**: `STRUCT`**Provider name**: `categoricalValue`**Description**: The categorical feature value.
          - `category_counts`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `categoryCounts`**Description**: Counts of all categories for the categorical feature. If there are more than ten categories, we return top ten (by count) and return one more CategoryCount with category "*OTHER*" and count as aggregate counts of remaining categories.
            - `category`**Type**: `STRING`**Provider name**: `category`**Description**: The name of category.
            - `count`**Type**: `INT64`**Provider name**: `count`**Description**: The count of training samples matching the category within the cluster.
        - `feature_column`**Type**: `STRING`**Provider name**: `featureColumn`**Description**: The feature column name.
        - `numerical_value`**Type**: `DOUBLE`**Provider name**: `numericalValue`**Description**: The numerical feature value. This is the centroid value for this feature.
    - `davies_bouldin_index`**Type**: `DOUBLE`**Provider name**: `daviesBouldinIndex`**Description**: Davies-Bouldin index.
    - `mean_squared_distance`**Type**: `DOUBLE`**Provider name**: `meanSquaredDistance`**Description**: Mean of squared distances between each sample to its cluster centroid.
  - `dimensionality_reduction_metrics`**Type**: `STRUCT`**Provider name**: `dimensionalityReductionMetrics`**Description**: Evaluation metrics when the model is a dimensionality reduction model, which currently includes PCA.
    - `total_explained_variance_ratio`**Type**: `DOUBLE`**Provider name**: `totalExplainedVarianceRatio`**Description**: Total percentage of variance explained by the selected principal components.
  - `multi_class_classification_metrics`**Type**: `STRUCT`**Provider name**: `multiClassClassificationMetrics`**Description**: Populated for multi-class classification/classifier models.
    - `aggregate_classification_metrics`**Type**: `STRUCT`**Provider name**: `aggregateClassificationMetrics`**Description**: Aggregate classification metrics.
      - `accuracy`**Type**: `DOUBLE`**Provider name**: `accuracy`**Description**: Accuracy is the fraction of predictions given the correct label. For multiclass this is a micro-averaged metric.
      - `f1_score`**Type**: `DOUBLE`**Provider name**: `f1Score`**Description**: The F1 score is an average of recall and precision. For multiclass this is a macro-averaged metric.
      - `log_loss`**Type**: `DOUBLE`**Provider name**: `logLoss`**Description**: Logarithmic Loss. For multiclass this is a macro-averaged metric.
      - `precision`**Type**: `DOUBLE`**Provider name**: `precision`**Description**: Precision is the fraction of actual positive predictions that had positive actual labels. For multiclass this is a macro-averaged metric treating each class as a binary classifier.
      - `recall`**Type**: `DOUBLE`**Provider name**: `recall`**Description**: Recall is the fraction of actual positive labels that were given a positive prediction. For multiclass this is a macro-averaged metric.
      - `roc_auc`**Type**: `DOUBLE`**Provider name**: `rocAuc`**Description**: Area Under a ROC Curve. For multiclass this is a macro-averaged metric.
      - `threshold`**Type**: `DOUBLE`**Provider name**: `threshold`**Description**: Threshold at which the metrics are computed. For binary classification models this is the positive class threshold. For multi-class classfication models this is the confidence threshold.
    - `confusion_matrix_list`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `confusionMatrixList`**Description**: Confusion matrix at different thresholds.
      - `confidence_threshold`**Type**: `DOUBLE`**Provider name**: `confidenceThreshold`**Description**: Confidence threshold used when computing the entries of the confusion matrix.
      - `rows`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `rows`**Description**: One row per actual label.
        - `actual_label`**Type**: `STRING`**Provider name**: `actualLabel`**Description**: The original label of this row.
        - `entries`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `entries`**Description**: Info describing predicted label distribution.
          - `item_count`**Type**: `INT64`**Provider name**: `itemCount`**Description**: Number of items being predicted as this label.
          - `predicted_label`**Type**: `STRING`**Provider name**: `predictedLabel`**Description**: The predicted label. For confidence_threshold > 0, we will also add an entry indicating the number of items under the confidence threshold.
  - `ranking_metrics`**Type**: `STRUCT`**Provider name**: `rankingMetrics`**Description**: Populated for implicit feedback type matrix factorization models.
    - `average_rank`**Type**: `DOUBLE`**Provider name**: `averageRank`**Description**: Determines the goodness of a ranking by computing the percentile rank from the predicted confidence and dividing it by the original rank.
    - `mean_average_precision`**Type**: `DOUBLE`**Provider name**: `meanAveragePrecision`**Description**: Calculates a precision per user for all the items by ranking them and then averages all the precisions across all the users.
    - `mean_squared_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredError`**Description**: Similar to the mean squared error computed in regression and explicit recommendation models except instead of computing the rating directly, the output from evaluate is computed against a preference which is 1 or 0 depending on if the rating exists or not.
    - `normalized_discounted_cumulative_gain`**Type**: `DOUBLE`**Provider name**: `normalizedDiscountedCumulativeGain`**Description**: A metric to determine the goodness of a ranking calculated from the predicted confidence by comparing it to an ideal rank measured by the original ratings.
  - `regression_metrics`**Type**: `STRUCT`**Provider name**: `regressionMetrics`**Description**: Populated for regression models and explicit feedback type matrix factorization models.
    - `mean_absolute_error`**Type**: `DOUBLE`**Provider name**: `meanAbsoluteError`**Description**: Mean absolute error.
    - `mean_squared_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredError`**Description**: Mean squared error.
    - `mean_squared_log_error`**Type**: `DOUBLE`**Provider name**: `meanSquaredLogError`**Description**: Mean squared log error.
    - `median_absolute_error`**Type**: `DOUBLE`**Provider name**: `medianAbsoluteError`**Description**: Median absolute error.
    - `r_squared`**Type**: `DOUBLE`**Provider name**: `rSquared`**Description**: R^2 score. This corresponds to r2_score in ML.EVALUATE.
- `model_level_global_explanation`**Type**: `STRUCT`**Provider name**: `modelLevelGlobalExplanation`**Description**: Output only. Global explanation contains the explanation of top features on the model level. Applies to both regression and classification models.
  - `class_label`**Type**: `STRING`**Provider name**: `classLabel`**Description**: Class label for this set of global explanations. Will be empty/null for binary logistic and linear regression models. Sorted alphabetically in descending order.
  - `explanations`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `explanations`**Description**: A list of the top global explanations. Sorted by absolute value of attribution in descending order.
    - `attribution`**Type**: `DOUBLE`**Provider name**: `attribution`**Description**: Attribution of feature.
    - `feature_name`**Type**: `STRING`**Provider name**: `featureName`**Description**: The full feature name. For non-numerical features, will be formatted like `.`. Overall size of feature name will always be truncated to first 120 characters.
- `results`**Type**: `UNORDERED_LIST_STRUCT`**Provider name**: `results`**Description**: Output only. Output of each iteration run, results.size() <= max_iterations.
  - `duration_ms`**Type**: `INT64`**Provider name**: `durationMs`**Description**: Time taken to run the iteration in milliseconds.
  - `eval_loss`**Type**: `DOUBLE`**Provider name**: `evalLoss`**Description**: Loss computed on the eval data at the end of iteration.
  - `index`**Type**: `INT32`**Provider name**: `index`**Description**: Index of the iteration, 0 based.
  - `learn_rate`**Type**: `DOUBLE`**Provider name**: `learnRate`**Description**: Learn rate used for this iteration.
  - `training_loss`**Type**: `DOUBLE`**Provider name**: `trainingLoss`**Description**: Loss computed on the training data at the end of iteration.
- `start_time`**Type**: `TIMESTAMP`**Provider name**: `startTime`**Description**: Output only. The start time of this training run.
- `training_options`**Type**: `STRUCT`**Provider name**: `trainingOptions`**Description**: Output only. Options that were used for this training run, includes user specified and default options that were used.
  - `adjust_step_changes`**Type**: `BOOLEAN`**Provider name**: `adjustStepChanges`**Description**: If true, detect step changes and make data adjustment in the input time series.
  - `approx_global_feature_contrib`**Type**: `BOOLEAN`**Provider name**: `approxGlobalFeatureContrib`**Description**: Whether to use approximate feature contribution method in XGBoost model explanation for global explain.
  - `auto_arima`**Type**: `BOOLEAN`**Provider name**: `autoArima`**Description**: Whether to enable auto ARIMA or not.
  - `auto_arima_max_order`**Type**: `INT64`**Provider name**: `autoArimaMaxOrder`**Description**: The max value of the sum of non-seasonal p and q.
  - `auto_arima_min_order`**Type**: `INT64`**Provider name**: `autoArimaMinOrder`**Description**: The min value of the sum of non-seasonal p and q.
  - `batch_size`**Type**: `INT64`**Provider name**: `batchSize`**Description**: Batch size for dnn models.
  - `booster_type`**Type**: `STRING`**Provider name**: `boosterType`**Description**: Booster type for boosted tree models.**Possible values**:
    - `BOOSTER_TYPE_UNSPECIFIED` - Unspecified booster type.
    - `GBTREE` - Gbtree booster.
    - `DART` - Dart booster.
  - `calculate_p_values`**Type**: `BOOLEAN`**Provider name**: `calculatePValues`**Description**: Whether or not p-value test should be computed for this model. Only available for linear and logistic regression models.
  - `clean_spikes_and_dips`**Type**: `BOOLEAN`**Provider name**: `cleanSpikesAndDips`**Description**: If true, clean spikes and dips in the input time series.
  - `color_space`**Type**: `STRING`**Provider name**: `colorSpace`**Description**: Enums for color space, used for processing images in Object Table. See more details at [https://www.tensorflow.org/io/tutorials/colorspace](https://www.tensorflow.org/io/tutorials/colorspace).**Possible values**:
    - `COLOR_SPACE_UNSPECIFIED` - Unspecified color space
    - `RGB` - RGB
    - `HSV` - HSV
    - `YIQ` - YIQ
    - `YUV` - YUV
    - `GRAYSCALE` - GRAYSCALE
  - `colsample_bylevel`**Type**: `DOUBLE`**Provider name**: `colsampleBylevel`**Description**: Subsample ratio of columns for each level for boosted tree models.
  - `colsample_bynode`**Type**: `DOUBLE`**Provider name**: `colsampleBynode`**Description**: Subsample ratio of columns for each node(split) for boosted tree models.
  - `colsample_bytree`**Type**: `DOUBLE`**Provider name**: `colsampleBytree`**Description**: Subsample ratio of columns when constructing each tree for boosted tree models.
  - `dart_normalize_type`**Type**: `STRING`**Provider name**: `dartNormalizeType`**Description**: Type of normalization algorithm for boosted tree models using dart booster.**Possible values**:
    - `DART_NORMALIZE_TYPE_UNSPECIFIED` - Unspecified dart normalize type.
    - `TREE` - New trees have the same weight of each of dropped trees.
    - `FOREST` - New trees have the same weight of sum of dropped trees.
  - `data_frequency`**Type**: `STRING`**Provider name**: `dataFrequency`**Description**: The data frequency of a time series.**Possible values**:
    - `DATA_FREQUENCY_UNSPECIFIED`
    - `AUTO_FREQUENCY` - Automatically inferred from timestamps.
    - `YEARLY` - Yearly data.
    - `QUARTERLY` - Quarterly data.
    - `MONTHLY` - Monthly data.
    - `WEEKLY` - Weekly data.
    - `DAILY` - Daily data.
    - `HOURLY` - Hourly data.
    - `PER_MINUTE` - Per-minute data.
  - `data_split_column`**Type**: `STRING`**Provider name**: `dataSplitColumn`**Description**: The column to split data with. This column won't be used as a feature. 1. When data_split_method is CUSTOM, the corresponding column should be boolean. The rows with true value tag are eval data, and the false are training data. 2. When data_split_method is SEQ, the first DATA_SPLIT_EVAL_FRACTION rows (from smallest to largest) in the corresponding column are used as training data, and the rest are eval data. It respects the order in Orderable data types: [https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#data-type-properties](https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#data-type-properties)
  - `data_split_eval_fraction`**Type**: `DOUBLE`**Provider name**: `dataSplitEvalFraction`**Description**: The fraction of evaluation data over the whole input data. The rest of data will be used as training data. The format should be double. Accurate to two decimal places. Default value is 0.2.
  - `data_split_method`**Type**: `STRING`**Provider name**: `dataSplitMethod`**Description**: The data split type for training and evaluation, e.g. RANDOM.**Possible values**:
    - `DATA_SPLIT_METHOD_UNSPECIFIED`
    - `RANDOM` - Splits data randomly.
    - `CUSTOM` - Splits data with the user provided tags.
    - `SEQUENTIAL` - Splits data sequentially.
    - `NO_SPLIT` - Data split will be skipped.
    - `AUTO_SPLIT` - Splits data automatically: Uses NO_SPLIT if the data size is small. Otherwise uses RANDOM.
  - `decompose_time_series`**Type**: `BOOLEAN`**Provider name**: `decomposeTimeSeries`**Description**: If true, perform decompose time series and save the results.
  - `distance_type`**Type**: `STRING`**Provider name**: `distanceType`**Description**: Distance type for clustering models.**Possible values**:
    - `DISTANCE_TYPE_UNSPECIFIED`
    - `EUCLIDEAN` - Eculidean distance.
    - `COSINE` - Cosine distance.
  - `dropout`**Type**: `DOUBLE`**Provider name**: `dropout`**Description**: Dropout probability for dnn models.
  - `early_stop`**Type**: `BOOLEAN`**Provider name**: `earlyStop`**Description**: Whether to stop early when the loss doesn't improve significantly any more (compared to min_relative_progress). Used only for iterative training algorithms.
  - `enable_global_explain`**Type**: `BOOLEAN`**Provider name**: `enableGlobalExplain`**Description**: If true, enable global explanation during training.
  - `feedback_type`**Type**: `STRING`**Provider name**: `feedbackType`**Description**: Feedback type that specifies which algorithm to run for matrix factorization.**Possible values**:
    - `FEEDBACK_TYPE_UNSPECIFIED`
    - `IMPLICIT` - Use weighted-als for implicit feedback problems.
    - `EXPLICIT` - Use nonweighted-als for explicit feedback problems.
  - `hidden_units`**Type**: `UNORDERED_LIST_INT64`**Provider name**: `hiddenUnits`**Description**: Hidden units for dnn models.
  - `holiday_region`**Type**: `STRING`**Provider name**: `holidayRegion`**Description**: The geographical region based on which the holidays are considered in time series modeling. If a valid value is specified, then holiday effects modeling is enabled.**Possible values**:
    - `HOLIDAY_REGION_UNSPECIFIED` - Holiday region unspecified.
    - `GLOBAL` - Global.
    - `NA` - North America.
    - `JAPAC` - Japan and Asia Pacific: Korea, Greater China, India, Australia, and New Zealand.
    - `EMEA` - Europe, the Middle East and Africa.
    - `LAC` - Latin America and the Caribbean.
    - `AE` - United Arab Emirates
    - `AR` - Argentina
    - `AT` - Austria
    - `AU` - Australia
    - `BE` - Belgium
    - `BR` - Brazil
    - `CA` - Canada
    - `CH` - Switzerland
    - `CL` - Chile
    - `CN` - China
    - `CO` - Colombia
    - `CS` - Czechoslovakia
    - `CZ` - Czech Republic
    - `DE` - Germany
    - `DK` - Denmark
    - `DZ` - Algeria
    - `EC` - Ecuador
    - `EE` - Estonia
    - `EG` - Egypt
    - `ES` - Spain
    - `FI` - Finland
    - `FR` - France
    - `GB` - Great Britain (United Kingdom)
    - `GR` - Greece
    - `HK` - Hong Kong
    - `HU` - Hungary
    - `ID` - Indonesia
    - `IE` - Ireland
    - `IL` - Israel
    - `IN` - India
    - `IR` - Iran
    - `IT` - Italy
    - `JP` - Japan
    - `KR` - Korea (South)
    - `LV` - Latvia
    - `MA` - Morocco
    - `MX` - Mexico
    - `MY` - Malaysia
    - `NG` - Nigeria
    - `NL` - Netherlands
    - `NO` - Norway
    - `NZ` - New Zealand
    - `PE` - Peru
    - `PH` - Philippines
    - `PK` - Pakistan
    - `PL` - Poland
    - `PT` - Portugal
    - `RO` - Romania
    - `RS` - Serbia
    - `RU` - Russian Federation
    - `SA` - Saudi Arabia
    - `SE` - Sweden
    - `SG` - Singapore
    - `SI` - Slovenia
    - `SK` - Slovakia
    - `TH` - Thailand
    - `TR` - Turkey
    - `TW` - Taiwan
    - `UA` - Ukraine
    - `US` - United States
    - `VE` - Venezuela
    - `VN` - Viet Nam
    - `ZA` - South Africa
  - `horizon`**Type**: `INT64`**Provider name**: `horizon`**Description**: The number of periods ahead that need to be forecasted.
  - `hparam_tuning_objectives`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `hparamTuningObjectives`**Description**: The target evaluation metrics to optimize the hyperparameters for.
  - `include_drift`**Type**: `BOOLEAN`**Provider name**: `includeDrift`**Description**: Include drift when fitting an ARIMA model.
  - `initial_learn_rate`**Type**: `DOUBLE`**Provider name**: `initialLearnRate`**Description**: Specifies the initial learning rate for the line search learn rate strategy.
  - `input_label_columns`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `inputLabelColumns`**Description**: Name of input label columns in training data.
  - `instance_weight_column`**Type**: `STRING`**Provider name**: `instanceWeightColumn`**Description**: Name of the instance weight column for training data. This column isn't be used as a feature.
  - `integrated_gradients_num_steps`**Type**: `INT64`**Provider name**: `integratedGradientsNumSteps`**Description**: Number of integral steps for the integrated gradients explain method.
  - `item_column`**Type**: `STRING`**Provider name**: `itemColumn`**Description**: Item column specified for matrix factorization models.
  - `kmeans_initialization_column`**Type**: `STRING`**Provider name**: `kmeansInitializationColumn`**Description**: The column used to provide the initial centroids for kmeans algorithm when kmeans_initialization_method is CUSTOM.
  - `kmeans_initialization_method`**Type**: `STRING`**Provider name**: `kmeansInitializationMethod`**Description**: The method used to initialize the centroids for kmeans algorithm.**Possible values**:
    - `KMEANS_INITIALIZATION_METHOD_UNSPECIFIED` - Unspecified initialization method.
    - `RANDOM` - Initializes the centroids randomly.
    - `CUSTOM` - Initializes the centroids using data specified in kmeans_initialization_column.
    - `KMEANS_PLUS_PLUS` - Initializes with kmeans++.
  - `l1_regularization`**Type**: `DOUBLE`**Provider name**: `l1Regularization`**Description**: L1 regularization coefficient.
  - `l2_regularization`**Type**: `DOUBLE`**Provider name**: `l2Regularization`**Description**: L2 regularization coefficient.
  - `learn_rate`**Type**: `DOUBLE`**Provider name**: `learnRate`**Description**: Learning rate in training. Used only for iterative training algorithms.
  - `learn_rate_strategy`**Type**: `STRING`**Provider name**: `learnRateStrategy`**Description**: The strategy to determine learn rate for the current iteration.**Possible values**:
    - `LEARN_RATE_STRATEGY_UNSPECIFIED`
    - `LINE_SEARCH` - Use line search to determine learning rate.
    - `CONSTANT` - Use a constant learning rate.
  - `loss_type`**Type**: `STRING`**Provider name**: `lossType`**Description**: Type of loss function used during training run.**Possible values**:
    - `LOSS_TYPE_UNSPECIFIED`
    - `MEAN_SQUARED_LOSS` - Mean squared loss, used for linear regression.
    - `MEAN_LOG_LOSS` - Mean log loss, used for logistic regression.
  - `max_iterations`**Type**: `INT64`**Provider name**: `maxIterations`**Description**: The maximum number of iterations in training. Used only for iterative training algorithms.
  - `max_parallel_trials`**Type**: `INT64`**Provider name**: `maxParallelTrials`**Description**: Maximum number of trials to run in parallel.
  - `max_time_series_length`**Type**: `INT64`**Provider name**: `maxTimeSeriesLength`**Description**: Get truncated length by last n points in time series. Use separately from time_series_length_fraction and min_time_series_length.
  - `max_tree_depth`**Type**: `INT64`**Provider name**: `maxTreeDepth`**Description**: Maximum depth of a tree for boosted tree models.
  - `min_relative_progress`**Type**: `DOUBLE`**Provider name**: `minRelativeProgress`**Description**: When early_stop is true, stops training when accuracy improvement is less than 'min_relative_progress'. Used only for iterative training algorithms.
  - `min_split_loss`**Type**: `DOUBLE`**Provider name**: `minSplitLoss`**Description**: Minimum split loss for boosted tree models.
  - `min_time_series_length`**Type**: `INT64`**Provider name**: `minTimeSeriesLength`**Description**: Set fast trend ARIMA_PLUS model minimum training length. Use in pair with time_series_length_fraction.
  - `min_tree_child_weight`**Type**: `INT64`**Provider name**: `minTreeChildWeight`**Description**: Minimum sum of instance weight needed in a child for boosted tree models.
  - `model_uri`**Type**: `STRING`**Provider name**: `modelUri`**Description**: Google Cloud Storage URI from which the model was imported. Only applicable for imported models.
  - `non_seasonal_order`**Type**: `STRUCT`**Provider name**: `nonSeasonalOrder`**Description**: A specification of the non-seasonal part of the ARIMA model: the three components (p, d, q) are the AR order, the degree of differencing, and the MA order.
    - `d`**Type**: `INT64`**Provider name**: `d`**Description**: Order of the differencing part.
    - `p`**Type**: `INT64`**Provider name**: `p`**Description**: Order of the autoregressive part.
    - `q`**Type**: `INT64`**Provider name**: `q`**Description**: Order of the moving-average part.
  - `num_clusters`**Type**: `INT64`**Provider name**: `numClusters`**Description**: Number of clusters for clustering models.
  - `num_factors`**Type**: `INT64`**Provider name**: `numFactors`**Description**: Num factors specified for matrix factorization models.
  - `num_parallel_tree`**Type**: `INT64`**Provider name**: `numParallelTree`**Description**: Number of parallel trees constructed during each iteration for boosted tree models.
  - `num_trials`**Type**: `INT64`**Provider name**: `numTrials`**Description**: Number of trials to run this hyperparameter tuning job.
  - `optimization_strategy`**Type**: `STRING`**Provider name**: `optimizationStrategy`**Description**: Optimization strategy for training linear regression models.**Possible values**:
    - `OPTIMIZATION_STRATEGY_UNSPECIFIED`
    - `BATCH_GRADIENT_DESCENT` - Uses an iterative batch gradient descent algorithm.
    - `NORMAL_EQUATION` - Uses a normal equation to solve linear regression problem.
  - `sampled_shapley_num_paths`**Type**: `INT64`**Provider name**: `sampledShapleyNumPaths`**Description**: Number of paths for the sampled Shapley explain method.
  - `subsample`**Type**: `DOUBLE`**Provider name**: `subsample`**Description**: Subsample fraction of the training data to grow tree to prevent overfitting for boosted tree models.
  - `tf_version`**Type**: `STRING`**Provider name**: `tfVersion`**Description**: Based on the selected TF version, the corresponding docker image is used to train external models.
  - `time_series_data_column`**Type**: `STRING`**Provider name**: `timeSeriesDataColumn`**Description**: Column to be designated as time series data for ARIMA model.
  - `time_series_id_column`**Type**: `STRING`**Provider name**: `timeSeriesIdColumn`**Description**: The time series id column that was used during ARIMA model training.
  - `time_series_id_columns`**Type**: `UNORDERED_LIST_STRING`**Provider name**: `timeSeriesIdColumns`**Description**: The time series id columns that were used during ARIMA model training.
  - `time_series_length_fraction`**Type**: `DOUBLE`**Provider name**: `timeSeriesLengthFraction`**Description**: Get truncated length by fraction in time series.
  - `time_series_timestamp_column`**Type**: `STRING`**Provider name**: `timeSeriesTimestampColumn`**Description**: Column to be designated as time series timestamp for ARIMA model.
  - `tree_method`**Type**: `STRING`**Provider name**: `treeMethod`**Description**: Tree construction algorithm for boosted tree models.**Possible values**:
    - `TREE_METHOD_UNSPECIFIED` - Unspecified tree method.
    - `AUTO` - Use heuristic to choose the fastest method.
    - `EXACT` - Exact greedy algorithm.
    - `APPROX` - Approximate greedy algorithm using quantile sketch and gradient histogram.
    - `HIST` - Fast histogram optimized approximate greedy algorithm.
  - `trend_smoothing_window_size`**Type**: `INT64`**Provider name**: `trendSmoothingWindowSize`**Description**: The smoothing window size for the trend component of the time series.
  - `user_column`**Type**: `STRING`**Provider name**: `userColumn`**Description**: User column specified for matrix factorization models.
  - `wals_alpha`**Type**: `DOUBLE`**Provider name**: `walsAlpha`**Description**: Hyperparameter for matrix factoration when implicit feedback type is specified.
  - `warm_start`**Type**: `BOOLEAN`**Provider name**: `warmStart`**Description**: Whether to train a model from the last checkpoint.
  - `xgboost_version`**Type**: `STRING`**Provider name**: `xgboostVersion`**Description**: User-selected XGBoost versions for training of XGBoost models.
- `training_start_time`**Type**: `INT64`**Provider name**: `trainingStartTime`**Description**: Output only. The start time of this training run, in milliseconds since epoch.
- `vertex_ai_model_id`**Type**: `STRING`**Provider name**: `vertexAiModelId`**Description**: The model id in the [Vertex AI Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction) for this training run.
- `vertex_ai_model_version`**Type**: `STRING`**Provider name**: `vertexAiModelVersion`**Description**: Output only. The model version in the [Vertex AI Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction) for this training run.

## `zone_id`{% #zone_id %}

**Type**: `STRING`