Alongside regression, we'd also like to have classification analysis type where the output value is categorical rather than numerical.

Steps:

• classification analysis type in C++
  Introduce classification analysis runner. ml-cpp#701
• classification analysis type in Java
  Implement new analysis type: classification #46537
• evaluation API
  Allow evaluation to consist of multiple steps. #46653
  Implement evaluation API for multiclass classification problem #47126
• HLRC
  Implement new analysis type: classification #46537
  Implement evaluation API for multiclass classification problem #47126
• HLRC docs and doc tests
  Add MlClientDocumentationIT tests for classification. #47569

Enhancements:

• support for boolean and integer types for dependent variable
• set the default for num_top_classes param to 2
• make format of MulticlassConfusionMatrix metric result more understandable
• default "prediction_field_name" to (dependent_variable + "_prediction")