Recommender Systems

Overview


Recommender systems are machine learning systems that are designed to help find products that a given customer would find interesting or relevant. They are particularly useful when:

  • They help customers find products they would not have normally found on their own
  • They can reduce the menu of produts offered to a customer based on the cognitive bias of decision paralysis.

Architectures


  • Predicting how a Customer Would Rate a Product - this involves creating a rating system for the company's products, and then forecasting how the given customer would rate each product.
  • Creating an Ordering Among of Set of Products - instead of rating products, this type of system just orders them such that the top product is thought to be the product most preferred by the customer.

Topics


  • Rating Matrices