Machine Learning Fundamentals

  1. Supervised Learning: Classification
    1. Concept of Classification: Predicting Categorical Labels
      1. Use Cases and Examples
        1. Medical Diagnosis
          1. Image Recognition
            1. Text Classification
              1. Quality Control
            2. Types of Classification
              1. Binary Classification
                1. Two-Class Problems
                  1. Decision Boundaries
                    1. Threshold Selection
                    2. Multiclass Classification
                      1. Multiple Categories
                        1. One-vs-Rest Strategy
                          1. One-vs-One Strategy
                          2. Multi-label Classification
                            1. Multiple Labels per Instance
                              1. Label Dependencies
                                1. Evaluation Challenges
                              2. Common Classification Algorithms
                                1. Logistic Regression
                                  1. Model Equation
                                    1. Sigmoid Function
                                      1. Log-Odds Transformation
                                      2. Decision Boundary
                                        1. Linear Separability
                                          1. Probability Thresholds
                                          2. Maximum Likelihood Estimation
                                            1. Regularization Options
                                            2. k-Nearest Neighbors (k-NN)
                                              1. Distance Metrics
                                                1. Euclidean Distance
                                                  1. Manhattan Distance
                                                    1. Cosine Similarity
                                                      1. Custom Metrics
                                                      2. Choosing k
                                                        1. Odd vs Even Values
                                                          1. Cross-Validation Selection
                                                            1. Bias-Variance Trade-off
                                                            2. Weighted Voting
                                                              1. Computational Complexity
                                                              2. Support Vector Machines (SVMs)
                                                                1. Linear SVM
                                                                  1. Maximum Margin Principle
                                                                    1. Support Vectors
                                                                      1. Soft Margin
                                                                      2. Kernel Trick
                                                                        1. Polynomial Kernels
                                                                          1. RBF Kernels
                                                                            1. Custom Kernels
                                                                            2. Parameter Tuning
                                                                              1. C Parameter
                                                                                1. Gamma Parameter
                                                                              2. Decision Trees
                                                                                1. Tree Structure
                                                                                  1. Nodes and Leaves
                                                                                    1. Splitting Rules
                                                                                      1. Tree Depth
                                                                                      2. Splitting Criteria
                                                                                        1. Information Gain
                                                                                          1. Gini Impurity
                                                                                            1. Chi-Square
                                                                                            2. Pruning Techniques
                                                                                              1. Pre-Pruning
                                                                                                1. Post-Pruning
                                                                                                  1. Cost Complexity Pruning
                                                                                                  2. Handling Categorical Features
                                                                                                  3. Naive Bayes
                                                                                                    1. Probabilistic Model
                                                                                                      1. Bayes' Theorem
                                                                                                        1. Prior Probabilities
                                                                                                          1. Likelihood Estimation
                                                                                                          2. Assumptions
                                                                                                            1. Feature Independence
                                                                                                              1. Impact of Violations
                                                                                                              2. Variants
                                                                                                                1. Gaussian Naive Bayes
                                                                                                                  1. Multinomial Naive Bayes
                                                                                                                    1. Bernoulli Naive Bayes
                                                                                                                    2. Laplace Smoothing
                                                                                                                  2. Evaluating Classification Models
                                                                                                                    1. Accuracy
                                                                                                                      1. Overall Correctness
                                                                                                                        1. Limitations with Imbalanced Data
                                                                                                                        2. The Confusion Matrix
                                                                                                                          1. True Positives
                                                                                                                            1. True Negatives
                                                                                                                              1. False Positives
                                                                                                                                1. False Negatives
                                                                                                                                  1. Matrix Interpretation
                                                                                                                                  2. Precision
                                                                                                                                    1. Positive Predictive Value
                                                                                                                                      1. Type I Error Consideration
                                                                                                                                      2. Recall (Sensitivity)
                                                                                                                                        1. True Positive Rate
                                                                                                                                          1. Type II Error Consideration
                                                                                                                                          2. F1-Score
                                                                                                                                            1. Harmonic Mean
                                                                                                                                              1. Balanced Metric
                                                                                                                                                1. Weighted F1-Score
                                                                                                                                                2. Specificity
                                                                                                                                                  1. True Negative Rate
                                                                                                                                                    1. Complement to Sensitivity
                                                                                                                                                    2. ROC Curve and AUC Score
                                                                                                                                                      1. True Positive Rate vs False Positive Rate
                                                                                                                                                        1. Area Under Curve Interpretation
                                                                                                                                                          1. Threshold Independence
                                                                                                                                                          2. Precision-Recall Curve
                                                                                                                                                            1. Precision vs Recall Trade-off
                                                                                                                                                              1. Area Under PR Curve
                                                                                                                                                                1. Imbalanced Data Suitability
                                                                                                                                                                2. Class Imbalance Considerations
                                                                                                                                                                  1. Sampling Techniques
                                                                                                                                                                    1. Cost-Sensitive Learning
                                                                                                                                                                      1. Threshold Adjustment

                                                                                                                                                                  Previous

                                                                                                                                                                  4. Supervised Learning: Regression

                                                                                                                                                                  Go to top

                                                                                                                                                                  Next

                                                                                                                                                                  6. Unsupervised Learning: Clustering