Data source: www.movielens.org

The goal of recommender systems is to provide personalized product recommendations to users. These systems can suggest items to purchase, shows or movies to watch, articles to read, and much more. Well known organizations that rely on predictive user modeling for personalization approaches are Netflix and Amazon. Netflix systems recommend shows and movies for users to watch, while Amazon recommends similar items for a user to purchase.

The approaches to personalized recommender systems that were deployed in this project were k-nearest-neighbor using labels derived from k-means clustering analysis, item-based collaborative filtering, user-based collaborative filtering and matrix factorization with singular value decomposition.

The purpose of this project is to perform and evaluate modeling techniques with the intention of building a recommender system based upon the lowest error results. This report provides the details of the data set, the various algorithms experimented, and the performance results that dictated the final recommender system our team decided to build.

• [Taihua Li] (https://www.linkedin.com/in/taihuali)
• [Chao Yu] (https://www.linkedin.com/in/charles-yu-4b046369)
• [Tammie Nguyen] (https://www.linkedin.com/in/tammie-nguyen-19061a12)

Note: this project is written in [Python 2.7] (https://www.python.org/download/releases/2.7/)