• Overview
• Algorithms
• Results
• Source code
• About

Source code and documentation

• Download the whole source code as a gzipped tarball [~465 KB]. Also available [~12 KB] is code that we used to parse the transcript data we received.
• Javadoc documentation for movies recommender systems
• Javadoc documentation for courses recommender systems

Starting Points

If you want to run our code, here are some good places to start:

• netflix.memreader.MemReader/transcript.memreader.TranscriptMemReader - These handle the creation of a memreader for either movie or transcript information.
• netflix.ui.RecommenderRunner - Starts a gui for either movie or transcript data.
• netflix.algorithms.modelbased.itembased.ItemBasedModelBuilder - A model builder for item-based collaborative filtering. MovieLensDBItemBasedModelBuilder and MovieLensMemItemBasedModelBuilder in the same package provide examples of how to use this class and other helper classes to build a model.