Feature preprocessing overview

Feature preprocessing is one of the most important steps in the machinelearning lifecycle. It consists of creating features and cleaning the trainingdata. Creating features is also referred asfeature engineering.

BigQuery ML provides the following feature preprocessing techniques:

• Automatic preprocessing. BigQuery ML performs automaticpreprocessing during training. For more information, seeAutomatic featurepreprocessing.

• Manual preprocessing. You can use theTRANSFORM clausein theCREATE MODEL statement to define custom preprocessing usingmanualpreprocessingfunctions.You can also use these functions outside of theTRANSFORM clause toprocess training data before creating the model.

Get feature information

You can use theML.FEATURE_INFOfunction toretrieve the statistics of all input feature columns.

Recommended knowledge

By using the default settings in theCREATE MODEL statements and theinference functions, you can create and use BigQuery ML modelseven without much ML knowledge. However, having basic knowledge about theML development lifecycle, such as feature engineering and model training,helps you optimize both your data and your model todeliver better results. We recommend using the following resources to developfamiliarity with ML techniques and processes:

• Machine Learning Crash Course
• Intro to Machine Learning
• Data Cleaning
• Feature Engineering
• Intermediate Machine Learning

What's next

• Learn aboutfeature serving inBigQuery ML.

• For more information about supported SQL statements and functions for modelsthat support feature preprocessing, see the following documents:

  • End-to-end user journeys for ML models
  • End-to-end user journeys for time series forecasting models
  • Contribution analysis user journey