Alternating Least Squares (ALS) with PySpark

Large Scale Computing: University of Genoa (101799) - Fall 2019 - Final project

Live Notebook on Colab

Abstract

In this project, I will present Alternating Least Squares (ALS) algorithm in the contest of modern recommender systems. Here is the list of content:

• Recommender system overview with emphasis on collaborative filtering
• Description of the ALS algorithm
• Implementation of the ALS algorithm from scratch: sequential (ALS_local.py) and parallel version (ALS_parallel.py using Spark)
• Usage of ALS provided by Spark MLlib APIs to predict the ratings for the movies in MovieLens Datasets