Using error-correcting dependencies for collaborative filtering

Collaborative filtering aims to automate the process of organizing and recommending information to users. This process consists of predicting the user rating of a given item based on other users' ratings. We propose a new algorithm for tackling this problem based on discovering the functional error-correcting dependencies in a dataset by using the fractal dimension. We experimentally evaluate our algorithm and compare it to some of the baseline schemes. The experimental results presented in this paper prove that our approach improves the accuracy and the performance of the filtering. (c) 2008 Elsevier B.V. All rights reserved.