Class 6: Spark ALS based recommender

What you should do before the class

1. Develop a recommendation engine using Retailrocket
2. Evaluate your model performance with P@K, R@K, and NDCG.

Class content

1. Presentation: Present your result of the Retailroket data.
2. Excercise: discuss how to evaluate familiarity, novelty, and diversity. (Optional) Code the function in python or spark.
3. Discussion: How can you improve your model.

Resources

Spark fundamentals

A good repo to reference

Complete Guide on DataFrame Operations in PySpark

Spark ALS code

Spark ALS sample code

Microsoft Recommender Repo'

Spark modeling package

spark-sklearn

Two sigma time series model "Flint"