AIExplainer
Machine Learning Intermediate 2 min read

What is variable importances?

A measure of how much each variable contributes to a model's predictions

Variable importances are a way to understand which input variables are most important for a machine learning model to make accurate predictions. It helps identify the variables that have the most impact on the model's output.

Think of variable importances like a recipe for your favorite dish. Just as some ingredients are more crucial to the final taste than others, variable importances show which variables are the key ingredients in your model's predictions.

In a model that predicts house prices, variable importances might show that the number of bedrooms and square footage are more important than the color of the walls or the type of flooring.

Variable importances are used to select the most relevant variables, reduce dimensionality, and improve model interpretability. They can also help identify correlations and relationships between variables.

A common misconception is that variable importances are the same as correlation coefficients. However, variable importances take into account the interactions between variables and their impact on the model's predictions.

The concept of variable importances originated in the field of decision trees and random forests, where it was used to measure the contribution of each variable to the model's predictions.

feature importances attribute importances predictor importances

Three products for different needs — explore what’s relevant to you.