Machine Learning syllabus

Derived reading outline. Source text split at semicolons, line breaks and sentence boundaries, not an official topic hierarchy.

• Introduction to machine learning, scope of machine learning, AI vs ML vs Data Science, traditional programming vs ML paradigm, and real-world engineering applications.
• Types of Learning: Supervised, unsupervised, semi-supervised, and reinforcement learning.
• Models of Machine Learning: Geometric model, probabilistic models, logical models, grouping and grading models, parametric and non-parametric models.
• Introduction to Feature Engineering.
• Feature Transformation: Dimensionality reduction techniques- Principal Component Analysis (PCA)
• Linear Discriminant Analysis (LDA).
• Case Study: Machine Learning Based Student Performance Prediction and Feature Engineering Analysis

Derived reading outline. Source text split at semicolons, line breaks and sentence boundaries, not an official topic hierarchy.

• Introduction to regression, need of regression, regression vs correlation.
• Types of regression: Univariate vs Multivariate, Linear vs Nonlinear, Simple vs Multiple, Bias–Variance Tradeoff, Overfitting and Underfitting.
• Regression Techniques: Simple and Multiple Linear Regression
• Polynomial Regression
• Decision Tree Regression, Random Forest Regression, Support Vector Regression.
• Regularization Techniques: Ridge Regression (L2)
• Lasso Regression (L1).
• Evaluation Metrics: Mean Squared Error (MSE)
• Mean Absolute Error (MAE)
• Root Mean Squared Error (RMSE)
• R-squared (R²).
• Case Study: Comparative Study of Regression Techniques for House Price Prediction

Derived reading outline. Source text split at semicolons, line breaks and sentence boundaries, not an official topic hierarchy.

• Introduction to classification, need of classification.
• Types of Classification: Binary and Multiclass, Binary vs. Multiclass Classification, Balanced and Imbalanced Classification Problems.
• Binary Classification: Linear classification model, decision boundary.
• Performance Evaluation: Confusion Matrix, Accuracy, Precision, Recall, F1-Score.
• Multiclass Classification: One-vs-One and One-vs-All classification techniques, multiclass confusion matrix
• Per-Class Precision and Per-Class Recall
• Macro, Micro and Weighted Averaging Methods.
• Classification Algorithms: K-Nearest Neighbors (KNN), Linear Support Vector Machine (SVM), Soft Margin SVM.
• Kernel Functions in SVM: Radial Basis Function (RBF/Gaussian) Kernel, Polynomial Kernel, Sigmoid Kernel.
• Case Study: Comparative Study of Classification Algorithms for Email Spam Detection.

Derived reading outline. Source text split at semicolons, line breaks and sentence boundaries, not an official topic hierarchy.

• Introduction to clustering, need for clustering, types of clustering, Hierarchical Clustering – Agglomerative and Divisive methods, Partitioning Methods: K-Means clustering algorithm, advantages and limitations, Elbow method, Silhouette method
• K-Medoids, Density-Based Clustering: DBSCAN algorithm, working mechanism, advantages and limitations.
• Distribution-Based Clustering: Gaussian Mixture Model.
• Applications, introduction to Ensemble Learning, homogeneous and heterogeneous ensemble methods, advantages and limitations.
• Basic Ensemble Techniques: Voting (Max Voting, Averaging, Weighted Averaging).
• Advanced Ensemble Techniques: Bagging and Random Forest.
• Boosting: AdaBoost, Gradient Boosting , Stacking.
• Case Study: Customer Segmentation and Sales Prediction

Derived reading outline. Source text split at semicolons, line breaks and sentence boundaries, not an official topic hierarchy.

• Introduction, need of reinforcement learning, components of reinforcement learning, comparison with supervised and unsupervised learning, applications and challenges of reinforcement learning.
• Markov Decision Process: Markov property, elements of MDP, episodic and continuing tasks.
• Reinforcement Learning Framework: Policy, state value function, action value function, Bellman equation, optimal policy.
• Reinforcement Learning Algorithms: Exploration vs Exploitation, ε-greedy strategy, dynamic programming, Q-Learning algorithm and update rule, simple reinforcement learning for game playing- Tic-Tac-Toe.
• Case Study: Smart Traffic Signal Control using Q-Learning.