Interpret Groups

Synopsis

This operator allows you to see how numeric columns differ for a given category.

Description

Given a categorical column, this operator calculates the normalized average values for all numerical columns grouped by the selected column. This value shows how much a sub-group differentiates from the total average. The values can be read as percentage. A value greater than zero indicates that the values in this group are on average larger, while a value smaller than zero indicates smaller values. Missing values are ignored for calculating the average values. But if a sub-group only contains missing values the returned value is also missing. The same is true if the groups contains infinite values, then the average also can't be calculated.

Input

training

The input data.

Output

example set

The calculated average values for each category.

Parameters

Differentiator

The category column you want to differentiate by.

Text explanation

If selected, the results are additionally returned in a plain text sentence.