Develop and compare user-neighbor and item-neighbor collaborative filtering with attention to sparse feedback and evaluation design.
Move from the recommender ranking problem to user-based neighborhoods, weighted candidate scores, and item-based similarity. Trace assumptions in interaction matrices, consider overlap and exposure, and compare methods with popularity baselines using ranking quality, coverage, cold-start tests, and sensitivity checks. Assignments use tiny invented matrices and emphasize that offline scores do not prove user benefit, fairness, or causal product impact.
Explain how user-item interactions support user-neighbor and item-neighbor recommendation scores. Identify sparsity, cold-start, exposure, similarity, scaling, and privacy limitations in collaborative filtering. Plan comparative evaluation using baselines, ranking and coverage measures, sensitivity tests, and clear caveats.
Introductory statistics, matrix/table literacy, and familiarity with ranking or supervised-learning evaluation are helpful. Coding is optional; assignments can be solved on paper with invented data. Avoid identifiable interaction histories and treat offline evaluation as limited evidence, not proof of real-world benefit.
Learners with basic familiarity who are ready to build practical, independent skills
People who learn best through examples, guided lessons, and hands-on practice
Professionals, creators, and independent builders looking for a repeatable workflow
Lesson 1 is free
4 lessons · Intermediate · Full course.
This course uses CC BY licensed material attributed to its content provider or uploader. Lesson playback is available only through authenticated access after publication and licensing checks pass.