Attribute Importance

What it does

This node scores how strongly each attribute relates to the target, without fitting a model. It ranks the attributes by mutual information or an F-test and returns the ranking (rank, attribute, importance, p_value for the F-test, method), plus an optional bar chart on the Artifacts tab. Runs on a kernel.

Inputs

One table: any mix of numeric, string, boolean or categorical attribute columns, plus a target column. Rows with a null target are dropped. Numeric attributes have missing values filled with the median; non-numeric ones are label-encoded. Regression needs a numeric target.

Settings

Data — the attribute columns and the target. The target is excluded from the attributes if you pick it twice. Task is auto-detected (strings, booleans and low-cardinality integers become classification); set it yourself if the guess is wrong.

Scoring — mutual information captures non-linear relationships and is the default; the F-test / ANOVA only measures linear association but returns a p-value alongside each score. Normalizing divides by the top score so importances land in 0–1. Top N truncates the output.

Visualization — publish the ranked bar chart under a name you pick, or turn it off for table only.