BDBP-207: Machine Learning Laboratory

Instructor: Shyam Rajagopalan
Institute: Institute of Bioinformatics and Applied Biotechnology (IBAB), Bengaluru
Term: January – April 2025

Overview

This repository contains lab notes, exercises, and Python implementations for the Machine Learning Laboratory (BDBP-207) course. It spans foundational concepts to advanced topics in machine learning, including supervised and unsupervised learning, ensemble models, support vector machines, Bayesian learning, and explainable AI.

Lab Modules

Foundations

• Lab 1: Functions and Derivatives
• Lab 2: scikit-learn Basics – California Housing Dataset
• Lab 3: Linear Regression
• Lab 4: Gradient Descent and Normal Equations

Classification and Evaluation

• Lab 5: Logistic Regression and SGD
• Lab 6: k-Fold Cross-Validation and Model Selection
• Lab 7: Data Preprocessing – Normalization and Standardization
• Lab 8: Regularization and Encoding for Categorical Data

Decision Trees and Ensembles

• Lab 9: Decision Trees with scikit-learn
• Lab 10: Decision Tree Components – Entropy, Information Gain
• Lab 11: Decision Tree Classifier from Scratch
• Lab 12: Decision Tree Regressor from Scratch
• Lab 13: Bagging and Random Forest
• Lab 14: AdaBoost
• Lab 15: Gradient Boosting
• Lab 16: XGBoost

Advanced Topics

• Lab 17: Kernel Methods – Feature Mapping and Kernel Trick
• Lab 18: RBF Kernel and Support Vector Machines
• Lab 19: Evaluation Metrics – Accuracy, ROC, AUC
• Lab 20: Unsupervised Learning – PCA and Clustering
• Lab 21: K-Means Clustering from Scratch
• Lab 22: Hierarchical Clustering
• Lab 23: Generative Models and Joint Probability
• Lab 24: Bayesian Learning – Naive Bayes and Inference
• Lab 25: Multiclass Classification – CIFAR10
• Lab 26: Explainable AI with SHAP
• Lab 27: Linear Algebra and Optimization Theory
• Lab 28: Real-World Case Studies – Genomics and DNA/RNA Data
• Lab 29: Project Implementation and Presentation

Key Skills Acquired

• Implementation of ML models from scratch and using libraries
• Supervised and unsupervised learning techniques
• Gradient descent and optimization
• Model evaluation and cross-validation
• Decision trees and ensemble learning
• Support vector machines and kernel methods
• Bayesian models and inference
• Data preprocessing and encoding
• Explainable AI tools such as SHAP
• Working with biological datasets

Datasets Used

scikit-learn Datasets

• load_iris
• load_digits
• load_wine
• load_breast_cancer
• load_diabetes
• fetch_california_housing
• fetch_20newsgroups
• fetch_openml

External Datasets

• Gene expression data: https://archive.ics.uci.edu/dataset/401/gene+expression+cancer+rna+seq
• SMS spam detection: https://www.kaggle.com/datasets/vishakhdapat/sms-spam-detection-dataset
• Breast cancer: https://raw.githubusercontent.com/jbrownlee/Datasets/master/breast-cancer.csv
• Twitter sentiment: https://www.kaggle.com/code/langkilde/linear-svm-classification-of-sentiment-in-tweets

Prerequisites

• Python programming
• Linear algebra and calculus basics
• Familiarity with pandas, NumPy, matplotlib
• Jupyter Notebooks or any Python IDE

Maintainers

Maintained by:
Sharmishta G Big Data Biology Program, IBAB
GitHub: https://github.com/SharmishtaGanesh14

References

• ISLP: An Introduction to Statistical Learning
• Pattern Recognition and Machine Learning by C. Bishop
• scikit-learn documentation: https://scikit-learn.org/stable/
• SHAP: https://shap.readthedocs.io/en/latest/