Slope One Meets Neighborhood: Revisiting Slope One Predictor in Collaborative Filtering

Abstract

Collaborative filtering (CF) framework in recommendationis a very popular technique for providing personalized recommendation.Slope one predictor is a model based CF which has received good atten-tion from researchers and practitioners. In this paper, we revisit the slopeone predictor to incorporate strong features of neighbourhood based CFinto it for providing personalized recommendation to users. Preliminary results with two real world datasets are very promising. Proposed tech-nique outperforms original slope one and its performance is at par with a variant of slope one introduced recently.