Recommendation of new items based on indexing techniques
Jian Chen, Jian Jun Yin, Jin H. Huang · 2005
The amount of information in the World Wide Web is increasing far more quickly than our ability to process. Recommender systems apply knowledge discovery techniques to help people find what they really want. These techniques include collaborative filtering (CF), association rules discovery and Bayesian networks, etc. Unfortunately all of these approaches have an important drawback: items or pages which being added to a site recently cannot be found. This is generally referred to as the "new item problem". We introduce a general framework for solving this problem and present a single index structure x-features-tree for using heuristic information retrieval technique to find the right items for the right users.