ML-15: Bagging and Random Forest
Master ensemble learning: understand why combining trees reduces variance, learn Bootstrap aggregating (Bagging), build Random Forests with feature randomization, and evaluate models using Out-of-Bag error.
ML-14: Decision Trees — Interpretable Tree Models
Master decision tree classification: understand Gini impurity and entropy, learn the CART algorithm step-by-step, and apply pruning to build interpretable yet powerful models.
ML-13: Naive Bayes: The Power of Generative Models
Master Naive Bayes classification: understand generative vs discriminative models, apply Bayes' theorem with the naive independence assumption, and build powerful text classifiers for spam detection and beyond.
ML-12: SVM: Kernel Methods and Soft Margin
Go beyond linear SVM: master the kernel trick for nonlinear boundaries, understand RBF and polynomial kernels, and learn soft-margin SVM with the C parameter for handling real-world noisy data.
ML-11: SVM: Hard Margin and Duality
Master hard-margin SVM from scratch: learn why maximizing margins improves generalization, derive the optimization problem step-by-step, and understand the powerful dual formulation with Lagrangian and KKT conditions.
ML-10: Gradient Descent Optimization
How gradient descent finds optimal parameters by iteratively descending the loss landscape. Covers learning rate tuning, Batch GD vs SGD vs Mini-batch, and advanced optimizers like Adam.
ML-09: K-Nearest Neighbors (KNN)
Master the simplest yet powerful classification algorithm: understand distance metrics, the curse of dimensionality, optimal K selection, and implement KNN from scratch with stunning decision boundary visualizations.
ML-08: Logistic Regression Basics
Learn why linear regression fails for classification and how the sigmoid function transforms predictions into probabilities. Includes cross-entropy loss derivation, gradient descent implementation from scratch, and multi-class strategies.
ML-07: Nonlinear Regression and Regularization
From straight lines to curves: master polynomial features, understand the bias-variance tradeoff, and tame overfitting with Ridge regularization and cross-validation.
ML-06: Linear Regression
A deep dive into linear regression: understand why squared error works, derive the normal equation step-by-step, explore the geometric intuition behind projections, and implement from scratch.