The course is divided into 13 modules, progressing from basics to advanced topics. Each module includes

Course Structure

📍 Module 1: Introduction to Machine Learning

• What is AI, ML, DL?
• Types of ML (Supervised, Unsupervised, Reinforcement)
• ML Workflow (Data → Model → Evaluation → Deployment)
• Parametric vs Non-Parametric Algorithms
• Use Cases of ML in Real Life
• Assignment 1: Identify ML applications around you

📍 Module 2: Exploratory Data Analysis (EDA)

• Data Overview (info, shape, describe)
• Handling Missing Values
• Handling Outliers
• Handling Skewness
• Data Encoding (Label, One-Hot, Ordinal)
• Feature Scaling (Normalization, Standardization)
• Feature Engineering (Feature Creation & Selection)
• Correlation Analysis & Heatmaps
• Assignment 2: Perform EDA on a real dataset

📍 Module 3: Regression Models

Linear Regression

• Concept & Mathematics
• Linear Regression with OLS
• Linear Regression with SGD
• Assumptions of Linear Regression
• Polynomial Regression
• Regularization (Ridge, Lasso, ElasticNet)
• Performance Metrics (MAE, MSE, RMSE, R² Score)
• Cross Validation

Logistic Regression

• Maximum Likelihood Estimation
• Performance Metrics (Confusion Matrix, Precision, Recall, F1, ROC AUC)
• Bias-Variance Tradeoff
• Overfitting & Underfitting
• Assignment 3: Predict house prices & classify survival in Titanic dataset

📍 Module 4: Tree-based Models

Decision Trees

• Entropy, Information Gain, Gini Index
• CART vs CHAID
• Pruning Techniques
• Performance Metrics

Random Forest

• Bagging & Bootstrap Sampling
• Feature Importance
• OOB Score
• Hyperparameter Tuning
• Assignment 4: Build decision tree & random forest on Heart Disease dataset

📍 Module 5: Ensemble Techniques

• Bagging vs Boosting vs Stacking
• AdaBoost
• Gradient Boosting
• XGBoost
• LightGBM & CatBoost (Intro)
• Assignment 5: Classify churn prediction using boosting models

📍 Module 6: Nearest Neighbours (KNN)

• Concept of KNN
• Distance Metrics (Euclidean, Manhattan, Cosine)
• Choosing Optimal K (Elbow Method)
• KNN in Classification vs Regression
• Assignment 6: Implement KNN on Iris dataset

📍 Module 7: Support Vector Machines (SVM)

• Understanding Hyperplanes & Support Vectors
• Hard vs Soft Margin
• Kernels (Linear, Polynomial, RBF, Sigmoid)
• SVM for Regression (SVR)
• Multi-class SVM (One-vs-One, One-vs-Rest)
• Assignment 7: Implement SVM with kernels on Breast Cancer dataset

📍 Module 8: Naive Bayes

• Bayes Theorem
• Gaussian, Multinomial, Bernoulli NB
• Applications (Spam Filtering, Text Classification)
• Assignment 8: Spam/Ham email classification using Naive Bayes

📍 Module 9: Dimensionality Reduction

• Curse of Dimensionality
• PCA (Steps, Explained Variance)
• LDA (Linear Discriminant Analysis)
• t-SNE & UMAP (Intro)
• Assignment 9: Apply PCA on MNIST dataset

📍 Module 10: Clustering

K-Means Clustering

• Working of Algorithm
• Choosing K (Elbow Method, Silhouette Score)
• Distance Measures

Hierarchical Clustering

• Agglomerative vs Divisive
• Dendrograms
• Linkage Criteria

DBSCAN

• Density-based clustering
• Epsilon & MinPts
• Assignment 10: Apply clustering on customer segmentation dataset

📍 Module 11: Model Evaluation & Tuning

• Train-Test Split & Cross Validation
• Bias-Variance Tradeoff
• Hyperparameter Tuning (GridSearchCV, RandomSearchCV, Optuna)
• Feature Importance (SHAP, LIME)
• Assignment 11: Hyperparameter tuning on Random Forest & XGBoost

📍 Module 12: Advanced ML Topics

• Handling Imbalanced Data (SMOTE, ADASYN)
• Anomaly Detection (Isolation Forest, One-Class SVM)
• Time Series Forecasting (ARIMA, Prophet, LSTMs – Intro)
• Recommendation Systems (Collaborative, Content-based, Hybrid)
• Semi-Supervised Learning
• Reinforcement Learning (MDP, Q-Learning – Intro)
• Assignment 12: Build a recommendation engine using MovieLens dataset

📍 Module 13: Mini Projects (Hands-On)

Project 5: Loan Default Prediction (Ensemble)

Project 1: Predict House Prices (Regression)

Project 2: Heart Disease Prediction (Classification)

Project 3: Customer Segmentation (Clustering)

Project 4: Spam Email Classifier (Naive Bayes)