ML Algorithms

Modules in this track

• What is Machine Learning? - ML fundamentals and types of learning
• Linear Regression - From scratch implementation
• Logistic Regression - Binary classification
• Regularization (L1/L2/ElasticNet) - Preventing overfitting by penalizing complexity
• Feature Scaling - Normalization and standardization for ML
• Model Evaluation Metrics - Precision, recall, F1, ROC-AUC
• Cross-Validation - K-Fold, stratified splits, reliable evaluation
• Naive Bayes - Probabilistic classifier
• K-Nearest Neighbors - Instance-based learning
• Decision Trees - Tree-based models
• Support Vector Machines - Maximum margin classifiers
• Multiclass Classification - One-vs-All, One-vs-One, softmax strategies
• Neural Networks - Backpropagation from scratch
• K-Means Clustering - Unsupervised learning
• Hierarchical Clustering - Dendrograms and linkage methods
• DBSCAN & Density Clustering - Density-based clustering
• PCA - Dimensionality reduction
• Ensemble Methods - Random forests, bagging
• Gradient Boosting - XGBoost, LightGBM, CatBoost
• Feature Importance & SHAP - Model interpretability and SHAP values
• Handling Imbalanced Data - SMOTE, class weights, resampling strategies
• Project: Algorithm Showdown - Pick the right algorithm for real-world medical diagnosis - compare 6+ models, evaluate with clinical metrics, and explain decisions with SHAP
• Project: IPL Match Predictor - End-to-end ML pipeline on 18 seasons of IPL data - 4-model calibrated ensemble, 20+ features, Monte Carlo validation