Highly accurate recommendation algorithm based on high-order similarities

摘要

In this Letter, we introduce a modified collaborative filtering (MCF)algorithm, which has remarkably higher accuracy than the standard collaborativefiltering. In the MCF, instead of the standard Pearson coefficient, theuser-user similarities are obtained by a diffusion process. Furthermore, byconsidering the second order similarities, we design an effective algorithmthat depresses the influence of mainstream preferences. The correspondingalgorithmic accuracy, measured by the ranking score, is further improved by24.9% in the optimal case. In addition, two significant criteria of algorithmicperformance, diversity and popularity, are also taken into account. Numericalresults show that the algorithm based on second order similarity can outperformthe MCF simultaneously in all three criteria.