📄️ Supervised learning
A complete study of supervised learning, from linear regression to ensemble methods. 7 hours, applied projects, a 40-question exam and a verifiable certificate.
📄️ 1. Framing a supervised problem
Module 1 of the Supervised Learning premium course: distinguishing regression from classification, defining the target and features, and building an honest dataset before even choosing a model.
📄️ 2. Linear regression and regularization
Module 2 of the Supervised Learning premium course: linear regression as a baseline, interpreting coefficients, and Ridge/Lasso regularization to control variance and select features.
📄️ 3. Logistic regression
Module 3 of the Supervised Learning premium course: turning a linear score into a probability with the sigmoid, understanding the decision boundary and threshold, and reading coefficients in classification.
📄️ 4. k-NN and SVM
Module 4 of the Supervised Learning premium course: the k-NN algorithm and the choice of k, margin and support vectors of the SVM, and the kernel trick for non-linear boundaries.
📄️ 5. Decision trees
Module 5 of the Supervised Learning premium course: how a tree splits the space with questions, impurity and choice of splits, readability of rules, and structural tendency to overfit.
📄️ 6. Random forests
Module 6 of the Supervised Learning premium course: bagging, double randomization of the random forest, out-of-bag score and feature importance - the robust variance remedy.
📄️ 7. Gradient boosting
Module 7 of the Supervised Learning premium course: the sequential logic of boosting, learning rate and key hyperparameters, XGBoost and LightGBM, and early stopping.
📄️ 8. Cross-validation and leakage
Module 8 of the Supervised Learning premium course: k-fold cross-validation, stratification, time-aware splits, and hunting data leakage with pipelines.
📄️ 9. Classification metrics
Module 9 of the Supervised Learning premium course: the confusion matrix, why accuracy misleads on imbalanced classes, the precision/recall trade-off, F1 and ROC AUC.
📄️ 10. Tuning and final project
Module 10 of the Supervised Learning premium course: grid search and random search, nested pipelines, and the complete workflow of a supervised project from framing to final model.
📄️ Recap and exam
Complete recap of the Supervised Learning premium course: models, validation and metrics module by module, then the 40-question certification exam.