DSBA Refresher HTML Index

Date: December 21, 2025 Author: P Baburaj Ambalam

Version 2.0 Comprehensive Q&A refreshers with examples and quizzes

Suggested Learning Path:
Foundation: Parametric vs Non-Parametric Models →
Tree-Based: Decision Trees → Random Forest → Boosting →
Engineering: Feature Engineering → ML Pipeline →
Clustering: K-means → Hierarchical/PCA

🎯 Foundational Concepts

Parametric vs Non-Parametric Models Foundational
Core concept: fixed functional forms vs data-driven structures; understanding model assumptions and capacity

Tree-Based Models

Decision Trees Beginner
Non-parametric models using recursive binary splits; interpretable but prone to overfitting
Decision Tree Data Analysis Guide Hands-On Tutorial
Step-by-step practical guide: data loading, cleaning, feature importance, pruning, hyperparameter tuning, and cross-validation
Bagging and Random Forests Intermediate
Bootstrap aggregating reduces variance; Random Forests add feature subsampling for decorrelation
Boosting Intermediate
Sequential weak learners correct errors; powerful but sensitive to noise and hyperparameters

🔧 ML Engineering & Pipelines

Feature Engineering & Cross Validation Beginner
Preprocessing, encoding, scaling, and validation strategies to prevent leakage and estimate generalization
ML Pipeline & Hyperparameter Tuning Intermediate
Reproducible pipelines, Grid/RandomizedSearchCV, and search space design for optimal configurations

🔍 Clustering & Dimensionality Reduction

K-means Clustering Beginner
Partition-based clustering minimizing WCSS; fast but assumes spherical clusters
Hierarchical Clustering & PCA Intermediate
Dendrogram-based clustering with linkage criteria; PCA for variance-preserving dimensionality reduction

About this collection: This collection starts with foundational concepts (parametric vs non-parametric models) then covers specific techniques. Each page includes technique descriptions at four levels (10-year-old, beginner, intermediate, expert), Q&A pairs, Python examples, 15-question quizzes with answers, practical checklists, common implementation errors, and authoritative references.