Boosting collaborative filtering based on statistical prediction errors
Shengchao Ding, Shiwan Zhao, Quan De Yuan, Xiatian Zhang, Rongyao Fu, Lawrence D. Bergman · 2008
User-based collaborative filtering methods typically predict a user's item ratings as a weighted average of the ratings given by similar users, where the weight is proportional to the user similarity. Therefore, the accuracy of user similarity is the key to the success of the recommendation, both for selecting neighborhoods and computing predictions. However, the computed similarities between users are somewhat inaccurate due to data sparsity.