BigQuery ML hyperparameter tuning overview

Introduction

In machine learning, hyperparameter tuning identifies a set of optimal hyperparameters for a learning algorithm. A hyperparameter is a model argument whose value is set before the learning process begins. By contrast, the values of other parameters such as coefficients of a linear model are learned.

Hyperparameter tuning lets you spend less time manually iterating hyperparameters and more time focusing on exploring insights from data.

For information about the supported model types for each SQL statement and function, and for a list of all of the supported SQL statements and functions for each model type, read the End-to-end user journey for each model.

Hyperparameter tuning offerings in BigQuery ML

BigQuery ML supports hyperparameter tuning for the following machine learning models, including time series and non-time series models:

Model types	Hyperparameter tuning details	Syntax
Linear & logistic regression, Kmeans, Matrix factorization, Boosted trees, Random forest, DNN, Wide-and-Deep	This hyperparameter tuning feature is made possible in BigQuery ML by using Vertex AI Vizier.	See Hyperparameter tuning for CREATE MODEL statements and ML.TRIAL_INFO.
AutoML Tables models	Automatic hyperparameter tuning is embedded in the AutoML Tables model training by default.	See CREATE MODEL statement for AutoML Tables models.
ARIMA_PLUS	The auto.ARIMA algorithm performs hyperparameter tuning for the trend module. Hyperparameter tuning is not supported for the entire modeling pipeline. See the modeling pipeline for more details.	See CREATE MODEL statement for time-series models and ML.ARIMA_EVALUATE.