Diving into BigQuery ML
Developing an ML model in BigQuery involves three main steps:
- Model creation, where you are required to choose the features and labels of your ML model and the options to tune the ML model. At this stage, BigQuery runs the training of the ML model on the training set that you've chosen.
- Model evaluation allows you to test the model trained in the previous step on a different set of records to prevent any overfitting.
- Model use: when the ML model is ready, we can apply it to a new dataset in order to make predictions or classifications of the labels according to the available features.
In the next paragraphs, we'll take a look at the syntax of these three stages and how these statements are built using stubs of code.
Creating the ML model (training)
When you've identified the ML use case and also the set of records to train your model, you can start training the model with the following query:
CREATE MODEL`<project_name...
