Machine Learning with Scikit-Learn

  1. Model Evaluation and Metrics
    1. Data Splitting Strategies
      1. Train-Test Split
        1. Purpose and Importance
          1. Split Ratios
            1. Random State
            2. Stratified Splitting
              1. Preserving Class Distribution
                1. Use in Classification
                2. Time Series Splitting
                  1. Temporal Data Considerations
                    1. TimeSeriesSplit
                    2. Group-based Splitting
                      1. GroupKFold
                        1. LeaveOneGroupOut
                      2. Cross-Validation
                        1. Purpose and Benefits
                          1. K-Fold Cross-Validation
                            1. Fold Selection
                              1. Computational Considerations
                              2. Stratified K-Fold
                                1. Class Balance Preservation
                                2. Leave-One-Out Cross-Validation
                                  1. Extreme Case of K-Fold
                                    1. Computational Cost
                                    2. Leave-P-Out Cross-Validation
                                      1. Shuffle Split
                                        1. Random Sampling
                                        2. Cross-Validation Utilities
                                          1. cross_val_score
                                            1. cross_validate
                                              1. cross_val_predict
                                            2. Classification Metrics
                                              1. Accuracy
                                                1. Definition and Calculation
                                                  1. Limitations with Imbalanced Data
                                                  2. Precision
                                                    1. True Positive Rate
                                                      1. Class-specific Precision
                                                      2. Recall
                                                        1. Sensitivity
                                                          1. Class-specific Recall
                                                          2. F1-Score
                                                            1. Harmonic Mean
                                                              1. Macro and Micro Averaging
                                                              2. F-beta Score
                                                                1. Weighted Harmonic Mean
                                                                2. Confusion Matrix
                                                                  1. True Positives
                                                                    1. False Positives
                                                                      1. True Negatives
                                                                        1. False Negatives
                                                                          1. Visualization
                                                                          2. Classification Report
                                                                            1. Comprehensive Metrics
                                                                            2. ROC Curve
                                                                              1. True Positive Rate vs False Positive Rate
                                                                                1. Threshold Selection
                                                                                  1. Multi-class Extension
                                                                                  2. AUC Score
                                                                                    1. Area Under the ROC Curve
                                                                                      1. Interpretation
                                                                                      2. Precision-Recall Curve
                                                                                        1. Imbalanced Dataset Evaluation
                                                                                        2. Average Precision Score
                                                                                          1. Log Loss
                                                                                            1. Probabilistic Predictions
                                                                                            2. Matthews Correlation Coefficient
                                                                                              1. Balanced Measure
                                                                                              2. Cohen's Kappa
                                                                                                1. Inter-rater Agreement
                                                                                                2. Hamming Loss
                                                                                                  1. Multi-label Classification
                                                                                                  2. Jaccard Score
                                                                                                    1. Set Similarity
                                                                                                  3. Regression Metrics
                                                                                                    1. Mean Absolute Error
                                                                                                      1. L1 Loss
                                                                                                        1. Robustness to Outliers
                                                                                                        2. Mean Squared Error
                                                                                                          1. L2 Loss
                                                                                                            1. Sensitivity to Outliers
                                                                                                            2. Root Mean Squared Error
                                                                                                              1. Interpretability
                                                                                                              2. Mean Absolute Percentage Error
                                                                                                                1. Relative Error
                                                                                                                2. R-squared
                                                                                                                  1. Coefficient of Determination
                                                                                                                    1. Explained Variance
                                                                                                                    2. Adjusted R-squared
                                                                                                                      1. Penalty for Additional Features
                                                                                                                      2. Mean Squared Log Error
                                                                                                                        1. Relative Errors
                                                                                                                        2. Median Absolute Error
                                                                                                                          1. Robust Metric
                                                                                                                          2. Explained Variance Score
                                                                                                                          3. Custom Scoring Functions
                                                                                                                            1. Creating Custom Metrics
                                                                                                                              1. make_scorer Function
                                                                                                                                1. Scorer Objects

                                                                                                                              Previous

                                                                                                                              6. Supervised Learning: Classification

                                                                                                                              Go to top

                                                                                                                              Next

                                                                                                                              8. Improving Model Performance