Intelligent Recommender System

Consider the following scenario. You tell the system what movies you have already seen and how you like them. The system uses this information to make suggestions in 2 ways:

• 1) collaborative filtering - it matches your tastes to those of the other users who share your likes and dislikes and suggests movies that they have seen and liked but you haven’t seen. (This is like a friend recommending a movie to you because some other friends, who are believed to be similar to you, like the movie.)

• 2) content-based approach - it recommends movies similar to those you prefer based on a comparison of movie content, i.e. according to the subject matter. (This is like a friend recommending a movie to you because he knows your preferences and have seen the movie.)

  The aim of this project is to develop a hybrid recommender system that incorporates the advantages of content-based and collaborative approaches while minimizing their shortcomings. The test domain is up to you - books, movies, CDs, e-newspaper articles, interesting web sites, restaurants recommendations.

  Further Reading

  1. D. Billsus, M. Pazzani (1998). Learning Collaborative Information Filters, Proc. of the 15th Int. Conf. on Machine Learning, San Francisco.
  2. R. Mooney, L.Roy (2000). Content-based Book Recommending Using Learning for Text Categorization, Proc. of the 5th ACM Conf. on Digital Libraries.
  See also Recommended Textbooks on Data Mining, Machine Learning, Neural Networks

  Here is an excellent site on Intelligent Recommender Systems.
  Freely avialable filtering toolkits.
  Collaborative Filtering.
  

  Task

  Visit the above sites. Find a freely available toolkit/sdk/framework which allows you to work in some aspect of recommender systems. For instance, the Stanford Information Filtering Tool includes two selective dissemination services, one for computer science technical reports and one for USENET news articles. The source code is also available. Papers describing the development of SIFT can be found here. A specialized filtering service run by the European Bioinformatics Institute which uses the SIFT package can be found here.

  A warm thanks to Irene Koprinska for coming up with this project idea.

  Home