E-commerce Clickstream data - find category user is interested in

by Tarandeep Singh   Last Updated July 12, 2019 09:19 AM - source

In an e-commerce site, there are N categories (shirts, trousers, jackets, coats...). From user's browsing history (most recent products viewed), how can I:

  • infer the category user is most interested in?
  • compute probability for above or associate some score with my inference
  • how to account for cases when user doesn't have enough history. How much history is enough to arrive at some reasonable conclusion?

Any help is greatly appreciated.

Related Questions

Difference between NMF, PMF and BMF

Updated April 12, 2017 14:19 PM

Matrix Factorization approaches

Updated April 18, 2017 10:19 AM