Spark Recommendation System

A simple recommendation system with ALS algorithm implemented in Apache Spark.

For this project, the movielens dataset has been imported into a MongoDB Atlas instance.

How does it work ?

From a userID value, we first print out the user preferences then infer some recommendations based on the user ratings.

An application_example.conf file is provided to set the database login values. Rename the application_example.conf to application.conf to use it in the application.

The final model will be updated after a grid search optimization.

Install

git clone https://github.com/Nohossat/spark_reco_system.git
cd spark_reco_system
sbt
compile
run

Improvements

• Allow the user to choose the userId from the CLI.
• Create an UI