Recommender Systems

13 April 2020

Enroll

Overview

Recommender systems had revolutionized the way we do online shopping. Contextual suggestions had proven to improve a company’s revenue by using subtle signals that are sometime hidden from the human observer. In this course we would cover various techniques of implementing and measuring a recommendation engine. We would discuss content-based approached vs collaborative based approaches, and demonstrate the trade offs. ​

Target Audience

Data scientists and technical product managers with coding skills.

Prerequisites

  • Familiarity with Jupyter Notebooks
  • Python Proficiency
  • Statistical Background is Recommended

Learning Outcomes

Graduates of this course will be able to:

  • Set up a recommendation engine
  • Measure the effectiveness of recommendation on a long term
  • Increase user retention with programmatic measure

Syllabus

Week
01

Content Based Recommendation

  • Knowledge representation
  • Human in the loop approach to recommendation
  • Measuring recommendation

Week
02

Collaborative Filtering

  • Scaling recommendation to millions of users
  • Association rule mining
  • Matrix factorization techniques
  • The cold-start problem

Week
03

Experimental Design

  • A-B testing for recommendation
  • Feedback loop
  • Avoiding user fatigue

Week
04

Deep Learning for Recommendations

  • Product2vec
  • Session modeling
  • Real-time recommendations

  • 4 Weeks
  • Sunday & Thursday
  • 18:00-20:00
  • Beginner
  • $50