How to Compare Machine Learning Models

This article describes how to compare the performance of multiple Machine Learning and Regression models by producing a table of metrics from each model as shown below.

Requirements

At least two Machine Learning models - for Method 1 only.
Outcome variable to be predicted (e.g., Prefer Cola) - for Method 2 only.
Predictors variables will be considered as predictors of the outcome variable (eg, Age, Gender, Exercise frequency) - for Method 2 only.

Note: Predictors that are considered to be uninformative will be automatically excluded from the model.

Method 1 - compare existing models

From the toolbar, go to Anything , or click + in the Report tree and select Advanced Analysis > Machine Learning > Compare Models.
In the object inspector , go to Data > Existing Models > Input Models and select the models you wish to compare.
Click Calculate if Calculate automatically is not ticked.

Method 2 - compare new models

From the toolbar, go to Anything , or click + from the Report tree, and select Advanced Analysis > Machine Learning > Compare Models.
In the object inspector , go to Data > Compare Machine Learning Models > Existing or new models, and select New models.
In Common Inputs > Outcome select the outcome variable to be predicted by the predictor variables.
Select the predictor variable(s) from the Predictor(s) list.
Go to Data > Model 1 > Algorithm and select the desired algorithm and settings.
Go to Data > Model 2 > Algorithm and select the desired algorithm and settings.
OPTIONAL: add additional models.