Projects: Module 03 — Supervised Learning¶

Project 1: Titanic Survival Predictor (Beginner)¶

Time estimate: 8–10 hours Use the Titanic dataset (https://www.kaggle.com/c/titanic). Engineer features (title from name, family size, fare per person). Compare at least 4 classifiers. Submit to Kaggle and document your approach.

Project 2: Salary Prediction Engine (Intermediate)¶

Time estimate: 10–14 hours Build a regression pipeline to predict salaries from job postings. Use the UCI Adult Income dataset. Handle categorical features with encoding. Use Lasso for feature selection. Explain the top 5 factors driving salary in plain language.

Project 3: Algorithm Benchmarking Suite (Advanced)¶

Time estimate: 12–16 hours Build a reusable benchmarking harness that: - Accepts any sklearn-compatible dataset - Trains all 6 algorithms from this module with default settings - Runs proper 5-fold stratified CV - Produces a comparison table with mean ± std for accuracy, AUC-ROC, and F1 - Runs on at least 5 different datasets and identifies which algorithm wins most often