fangshulin / recommender-system

Personalized music artist recommendation system with Spark

Music Artist Recommender System with Apache Spark

Built a personalized music artist recommendation system with Yahoo! music community ratings.

Data

I used Yahoo! Community music ratings data for this analysis. This Yahoo! Music data represents a sample of (anonymized) Yahoo! users' ratings of musical artists, gathered over a thirty-day period sometime prior to March 2004. The dataset contains 11,557,943 ratings of 98,211 artists by 1,948,882 anonymous users.

Steps

First I Performed online analytical processing (OLAP) based on Spark Dataframe and Spark SQL. Then Trained ALS model based on Collaborative Filtering in PySpark MLlib and tuned hyperparameters with cross-validation toolbox. The Spark API provides the implementation of the ALS algorithm, which is used to learn these latent factors based on the following six parameters:

• numBlocks: The number of blocks used to parallelise computation (set to -1 to auto-configure).
• rank: The number of latent factors in the model.
• iterations: The number of iterations of ALS to run. ALS typically converges to a reasonable solution in 20 iterations or less.
• lambda: The regularisation parameter.
• implicitPrefs: Whether to use the explicit feedback from the ALS variant for implicit feedback data.
• alpha: A parameter applicable to the implicit feedback variant of ALS that governs the baseline confidence in preference observations.