Machine Learning Module

An internal build 3.20.0.5.2 of the H2O Library is included with this release. This build fixes security vulnerabilities.

A new Configuration Settings tab has been added to the Machine Learning Model Management application.

Machine Learning Module settings are now configured on the Configuration Settings page. The page provides configuration for pool size, maximum and minimum memory settings, Java property settings, environment variables, and process arguments.

This release of the Machine Learning Module on the Linux platform includes the XGBOOST Classification and XGBOOST Regression stages as a technology preview.

The XGBOOST Classification and XGBOOST Regression stages are an early implementation containing fundamental capabilities and are likely to change in future releases. Using these stages in a production environment is not currently recommended. These stages are only supported on the Linux platform.