Feature Importance

Coefficient-based methods

linear regression, logistic regression)

Impurity-based methods

• Feature importance is calculated as the decrease in node impurity weighted by the probability of reaching that node.
• Tree-based models have a strong tendency to overestimate the importance of continuous numerical or high cardinality categorical features.

SHapley Additive exPlanations (SHAP)

• What is SHAP values:
  • SHAP values have their roots in cooperative game theory, where Shapley values are used to quantify the contribution of each player to the game.
• How does it work
  • It considers all possible combinations of features and measure the change in the model’s prediction when a specific feature is included or excluded.
• Why is it useful:
  • It explains predictions made by black-box machine learning models: given a prediction made by a machine learning model, SHAP values explain the prediction by quantifying the additive importance of each feature to the prediction.
  • The importance values can be positive or negative.
  • It can be calculated for each individual dataset to explore how features influenced the prediction result for this specific dataset
• Cons?
  • computationally expensive to compute SHAP values for general black-box models; in the case of trees and forests there exists a fast polynomial-time algorithm