Calculating Feature Importance in LightGBM Models

This function calculates the importance of each feature in a LightGBM model and returns a dictionary containing feature names and their importance scores.

Technology Stack : LightGBM, NumPy

Code Type : Function

Code Difficulty : Intermediate

                
                    
def random_feature_importance(lgbm_model, feature_names):
    """
    This function calculates the importance of each feature in a LightGBM model.
    """
    import lightgbm as lgb
    import numpy as np

    # Get the feature importance from the model
    importance = lgb_model.feature_importances_
    
    # Normalize the importance values
    normalized_importance = importance / np.sum(importance)
    
    # Sort the features based on their importance
    sorted_idx = np.argsort(normalized_importance)[::-1]
    
    # Create a feature importance dictionary
    feature_importance_dict = {feature_names[i]: normalized_importance[i] for i in sorted_idx}
    
    return feature_importance_dict