You are an online retailer/travel agent/movie review website, and you would like to help the visitors of your website to explore more of your products/destinations/movies. You got data which either describe the different products/destinations/films, or past transactions/trips/views (or preferences) of your visitors (or both!). You decide to leverage that data to provide relevant and meaningful recommendations.

During this tutorial, participants will learn how to apply unsupervised (content-based) and supervised (collaborative filtering) methods to provide product recommendations. As a use case, we will build a movie recommendation engine.

This tutorial is for people that

• know fundamentals of calculus and algorithmic
• are familiar with python, Numpy, Pandas

Curriculum

• Recommender systems landscape
• Focus on content-based and collaborative filtering (matrix factorization): principle and key aspects
• Build a movie recommendation engine (training, optimisation and evaluation). NB: Material will cover both content-based and collaborative filtering approaches; we will go deeper on one, we leave the audience decide which technique would be most useful for their particular use-case.
• Compare both approaches

Prerequisites

• Experience with Python and Jupyter notebooks
• Numpy, Pandas, Scikit-learn installed