Training the binary logistic regression model
As we already did in Chapter 4, Predicting Numerical Values with Linear Regression, we'll adopt an incremental approach in trying to improve the performance of our ML model at each attempt:
- Let's start training our first ML model,
binary_classification_version_1:CREATE OR REPLACE MODEL `05_chicago_taxi.binary_classification_version_1` OPTIONS (model_type='logistic_reg', labels = ['will_get_tip']) AS SELECT trip_seconds, IF(tips>0,1,0) AS will_get_tip FROM `05_chicago_taxi.training_table`;
In this BigQuery ML statement, we can see the
CREATE OR REPLACE MODELkeywords used to start the training of the model. These keywords are followed by the identifier of the ML model. After the identifier, we can notice theOPTIONSclause. As our options...