/recommenders-deconstructed

A repo for implementing, understanding and analyzing recommender systems in Python.

Primary LanguagePython

Recommenders Deconstructed

Description

A repo for implementing, understanding and analyzing recommender systems in Python.

Collaborative Filtering

Collaborative Filtering project using a subset of the millionsong dataset. User-User Collaborative Filtering implemented with three different ranking methods. Item-Item Collaborative Filtering implemented using several set and similarity methods:

  • Set Intersection
  • Ochiai
  • TF-IDF
  • Jaccard
  • Cosine
  • Smoothed Cosine

Matrix Factorization, Factorization Machines and Networks coming.

Requirements

  • Python
  • Pandas
  • Numpy