Automatic feature preprocessing

BigQuery ML performs automatic preprocessing during training by using the CREATE MODEL statement. Automatic preprocessing consists of missing value imputation and feature transformations.

For information about feature preprocessing support in BigQuery ML, see Feature preprocessing overview.

Missing data imputation

In statistics, imputation is used to replace missing data with substituted values. When you train a model in BigQuery ML, NULL values are treated as missing data. When you predict outcomes in BigQuery ML, missing values can occur when BigQuery ML encounters a NULL value or a previously unseen value. BigQuery ML handles missing data differently, based on the type of data in the column.

Column type	Imputation method
Numeric	In both training and prediction, `NULL` values in numeric columns are replaced with the mean value of the given column, as calculated by the feature column in the original input data.
One-hot/Multi-hot encoded	In both training and prediction, `NULL` values in the encoded columns are mapped to an additional category that is added to the data. Previously unseen data is assigned a weight of 0 during prediction.
`TIMESTAMP`	`TIMESTAMP` columns use a mixture of imputation methods from both standardized and one-hot encoded columns. For the generated Unix time column, BigQuery ML replaces values with the mean Unix time across the original columns. For other generated values, BigQuery ML assigns them to the respective `NULL` category for each extracted feature.
`STRUCT`	In both training and prediction, each field of the `STRUCT` is imputed according to its type.

Feature transformations

By default, BigQuery ML transforms input features as follows:

Input data type	Transformation method	Details
`INT64` `NUMERIC` `BIGNUMERIC` `FLOAT64`	Standardization	For most models, BigQuery ML standardizes and centers numerical columns at zero before passing it into training. The exceptions are boosted tree and random forest models, for which no standardization occurs, and k-means models, where the `STANDARDIZE_FEATURES` option controls whether numerical features are standardized.
`BOOL` `STRING` `BYTES` `DATE` `DATETIME` `TIME`	One-hot encoded	For all non-numerical, non-array columns other than `TIMESTAMP`, BigQuery ML performs a one-hot encoding transformation for all models other than boosted tree and random forest models. This transformation generates a separate feature for each unique value in the column. Label encoding transformation is applied to train boosted tree and random forest models to convert each unique value into a numerical value.
`ARRAY`	Multi-hot encoded	For all non-numerical `ARRAY` columns, BigQuery ML performs a multi-hot encoding transformation. This transformation generates a separate feature for each unique element in the `ARRAY`.
`TIMESTAMP`	Timestamp transformation	When a linear or logistic regression model encounters a `TIMESTAMP` column, it extracts a set of components from the `TIMESTAMP` and performs a mix of standardization and one-hot encoding on the extracted components. For the Unix epoch time in seconds component, BigQuery ML uses standardization. For all other components, it uses one-hot encoding. For more information, see the following timestamp feature transformation table.
`STRUCT`	Struct expansion	When BigQuery ML encounters a `STRUCT` column, it expands the fields inside the `STRUCT` to create a single column. It requires all fields of `STRUCT` to be named. Nested `STRUCT`s are not allowed. The column names after expansion are in the format of `{struct_name}_{field_name}`.
`ARRAY` of `STRUCT`	No transformation
`ARRAY` of `NUMERIC`	No transformation

`TIMESTAMP` feature transformation

The following table shows the components extracted from TIMESTAMP columns and the corresponding transformation method.

`TIMESTAMP` component	`processed_input` result	Transformation method
Unix epoch time in seconds	`[COLUMN_NAME]`	Standardization
Day of month	`_TS_DOM_[COLUMN_NAME]`	One-hot encoding
Day of week	`_TS_DOW_[COLUMN_NAME]`	One-hot encoding
Month of year	`_TS_MOY_[COLUMN_NAME]`	One-hot encoding
Hour of day	`_TS_HOD_[COLUMN_NAME]`	One-hot encoding
Minute of hour	`_TS_MOH_[COLUMN_NAME]`	One-hot encoding
Week of year (weeks begin on Sunday)	`_TS_WOY_[COLUMN_NAME]`	One-hot encoding
Year	`_TS_YEAR_[COLUMN_NAME]`	One-hot encoding

Category feature encoding

For features that are one-hot encoded, you can specify a different default encoding method by using the model option CATEGORY_ENCODING_METHOD. For generalized linear models (GLM) models, you can set CATEGORY_ENCODING_METHOD to one of the following values:

One-hot encoding

One-hot encoding maps each category that a feature has to its own binary feature, where 0 represents the absence of the feature and 1 represents the presence (known as a dummy variable). This mapping creates N new feature columns, where N is the number of unique categories for the feature across the training table.

For example, suppose your training table has a feature column that's called fruit with the categories Apple, Banana, and Cranberry, such as the following:

Row	fruit
1	Apple
2	Banana
3	Cranberry

In this case, the CATEGORY_ENCODING_METHOD='ONE_HOT_ENCODING' option transforms the table to the following internal representation:

Row	fruit_Apple	fruit_Banana	fruit_Cranberry
1	1	0	0
2	0	1	0
3	0	0	1

One-hot encoding is supported by linear and logistic regression and boosted tree models.

Dummy encoding

Dummy encoding is similar to one-hot encoding, where a categorical feature is transformed into a set of placeholder variables. Dummy encoding uses N-1 placeholder variables instead of N placeholder variables to represent N categories for a feature. For example, if you set CATEGORY_ENCODING_METHOD to 'DUMMY_ENCODING' for the same fruit feature column shown in the preceding one-hot encoding example, then the table is transformed to the following internal representation:

Row	fruit_Apple	fruit_Banana
1	1	0
2	0	1
3	0	0

The category with the most occurrences in the training dataset is dropped. When multiple categories have the most occurrences, a random category within that set is dropped.

The final set of weights from ML.WEIGHTS still includes the dropped category, but its weight is always 0.0. For ML.ADVANCED_WEIGHTS, the standard error and p-value for the dropped variable is NaN.

If warm_start is used on a model that was initially trained with 'DUMMY_ENCODING', the same placeholder variable is dropped from the first training run. Models cannot change encoding methods between training runs.

Dummy encoding is supported by linear and logistic regression models.

Label encoding

Label encoding transforms the value of a categorical feature to an INT64 value in [0, <number of categories>].

For example, if you had a book dataset like the following:

Title	Genre
Book 1	Fantasy
Book 2	Cooking
Book 3	History
Book 4	Cooking

The label encoded values might look similar to the following:

Title	Genre (text)	Genre (numeric)
Book 1	Fantasy	1
Book 2	Cooking	2
Book 3	History	3
Book 4	Cooking	2

The encoding vocabulary is sorted alphabetically. NULL values and categories that aren't in the vocabulary are encoded to 0.

Label encoding is supported by boosted tree models.

Target encoding

Target encoding replaces the categorical feature value with the probability of the target for classification models, or with the expected value of the target for regression models.

Features that have been target encoded might look similar to the following example:

 # Classification model +------------------------+----------------------+ | original value | target encoded value | +------------------------+----------------------+ | (category_1, target_1) | 0.5 | | (category_1, target_2) | 0.5 | | (category_2, target_1) | 0.0 | +------------------------+----------------------+ # Regression model +------------------------+----------------------+ | original value | target encoded value | +------------------------+----------------------+ | (category_1, 2) | 2.5 | | (category_1, 3) | 2.5 | | (category_2, 1) | 1.5 | | (category_2, 2) | 1.5 | +------------------------+----------------------+

Target encoding is supported by boosted tree models.

What's next

For more information about supported SQL statements and functions for models that support automatic feature preprocessing, see the following documents: