ECS 172: Recommender Systems

Subject
ECS 172
Title
Recommender Systems
Status
Active
Units
4.0
Effective Term
2020 Spring Quarter
Learning Activities
Lecture: 3 hours
Discussion: 1 hour
Description
Collaborative filtering and content-based methods for building recommender systems. Statistical, matrix factorization, textual analysis, and nearest-neighbor approaches. Case studies. GE Prior to Fall 2011: SciEng. GE: SE.
Prerequisites
(ECS 032B or ECS 036B or ECS 040); (ECS 132 or STA 130A or STA 131A or ECN 140); (MAT 022A or MAT 027A or MAT 067)
Enrollment Restrictions
Pass One restricted to Computer Science and Computer Science & Engineering students only.

 

Summary of Course Content:

  1. Review of probability concepts and linear algebra (both illustrated with R or NumPy).
  2. Collaborative filtering methods:
    Applied predictive modeling:  User's-view treatment of Machine Learning methods (linear/logistic regression, random forests, neural nets, nearest-neighbor); Matrix factorization models.
  3. Content-based recommendations
    Overview of text classification methods, including software; content-based systems
  4. Graphical models and methods
  5. Google PageRank
  6. Applications and case studies

Illustrative Reading:

Textbook:  Charu Aggarwal, "Recommender Systems: the Textbook," Springer, 2016, 498 pp., and/or instructor's notes; published case studies.

Potential Course Overlap:
About 5% of the material overlaps course 171. However, it is presented just as a brief overview, from a "user" perspective, without technical detail.  The situation is similar to the overlap between courses 174 and 171.

Final Exam:
Yes Final Exam

Course Category

Upper Division