Projects: Module 04 — Unsupervised Learning¶

Project 1: Customer Segmentation Dashboard (Intermediate)¶

Time estimate: 8–12 hours Use a retail dataset (e.g., UCI Online Retail). Compute RFM features. Apply k-means and DBSCAN. Compare segments. Write a business-readable report describing each segment and its marketing implications.

Project 2: Topic Modeling with NMF (Intermediate)¶

Time estimate: 8–10 hours Apply Non-negative Matrix Factorization to a text corpus (e.g., 20 Newsgroups dataset from sklearn). Identify latent topics. Compare to LDA (sklearn). Evaluate coherence.

Project 3: Anomaly Detection System (Advanced)¶

Time estimate: 12–16 hours Build an anomaly detection pipeline using Isolation Forest, One-Class SVM, and DBSCAN on the KDD Cup 1999 network intrusion dataset. Compare approaches, tune thresholds, and produce precision/recall metrics.