Unsupervised Learning and Practical ML

You need to sign in or sign up before continuing.

reading · we frame, you read MIT or the canonical taught · we author, no canonical fits ↺ spirals back to earlier lessons

Course locked

Complete Supervised Learning first.

This course unlocks once you've finished its prerequisite. Open prerequisite →

Prerequisites

Supervised Learning not yet

Complete prerequisites first to enroll.

reading · 35 min + ~230 min canonical

Learning Without Labels: Clustering and the Unsupervised Paradigm

Supervised learning had a luxury you might not have appreciated: the answers. Every training example came with its correct label, and learning was "fit a function to match the ans…

primary Andrew Ng ML Specialization — Course 3 on Coursera

preview →

reading · 35 min + ~230 min canonical L1

Dimensionality Reduction and Anomaly Detection

Real data is wide. A single customer record might have 200 columns; an image is thousands of pixels; a gene-expression sample has 20,000 features. And high-dimensional data is str…

primary Andrew Ng ML Specialization — Course 3 on Coursera

after L1 locked

reading · 35 min + ~230 min canonical L1

Recommender Systems: The Most Valuable Unsupervised Application

Here is arguably the single most commercially valuable application of machine learning ever built: the recommender system. The "Recommended for you" on every store, the "Up next" …

primary Andrew Ng ML Specialization — Course 3 on Coursera

after L2 locked

reading · 40 min + ~230 min canonical L1

The Practical ML Craft: Preprocessing, Tuning, and Interpretability

The algorithms get the glory, but the craft is what makes ML actually work — and this lesson is that craft: the unglamorous, high-leverage skills that separate a model that perfor…

primary Andrew Ng ML Specialization — Course 3 on Coursera

after L3 locked

reading · 40 min + ~230 min canonical L1

From Notebook to Production: MLOps and the ML Lifecycle

Here is what most ML courses get wrong: they end at model training. You build a model in a notebook, get a good accuracy score, and the course says "congratulations, you've learne…

primary Andrew Ng ML Specialization — Course 3 on Coursera

after L4 locked

exercise reading · 15 min L1

Exercises: Easy (Warm-up)

- [ ] Implement k-means clustering from scratch in Python — Random initialization, Lloyd's algorithm, plot iterations - [ ] Implement PCA from scratch and visualize — Reduce MNIST…

after L5 locked

exercise reading · 30 min L1

Exercises: Medium (Core)

- [ ] Implement DBSCAN from scratch in Python — Epsilon-neighborhood, core/border/noise points - [ ] Build a collaborative filtering recommender — User-item matrix, cosine similar…

after L6 locked

exercise reading · 45 min L1

Exercises: Hard (Stretch)

- [ ] Implement t-SNE from scratch in Python — Student-t distribution, gradient descent on KL divergence - [ ] Build a content-based recommender with TF-IDF — Cosine similarity on…

after L7 locked

exercise reading · 60 min L1

Mini-Project

- [ ] Build a customer segmentation system in Python — Load e-commerce data, feature engineering (RFM analysis), k-means + DBSCAN clustering, dimensionality reduction for visualiz…

after L8 locked

L10

project reading · 360 min L1

Project: Customer Segmentation

Load e-commerce data, perform RFM (Recency, Frequency, Monetary) feature engineering, apply k-means and DBSCAN clustering, reduce dimensions with PCA for visualization, and profil…

after L9 locked

L11

project reading · 300 min L1

Project: Recommender System

Build a movie recommender using collaborative filtering on the MovieLens 100K dataset. Implement user-based and item-based approaches with cosine similarity. Evaluate with RMSE on…

after L10 locked

L12

assessment reading · 56 min L1

Self-Assessment: Unsupervised Learning and Practical ML

- [ ] Explain k-means clustering. What are its limitations? - [ ] What is DBSCAN? How does it handle clusters of different shapes and sizes? - [ ] Explain PCA as dimensionality re…

after L11 locked

12 lessons. Read in order; spiral back when you need to. By the end you'll have used the core ideas twice — once on the abstract, once on something you'll meet at work next week.