An Effective Technique for Personalization Recommendation Based on Access Sequential Patterns

Abstract

Considering that personalization recommendation systems based on association rules suffer from some limitations that a lot of time is spent on matching current user session with all discovered patterns in patterns database, authors propose a new approach to build personalization recommendation system based on access sequential patterns discovered form usage data and highly compressed into a tree structure. During personalization recommendation stage we just need to intercept nearest access subsequence from current user session to match some sub paths of the tree. The speed of pattern matching is improved enormously, which satisfies the need of real-time recommendation better. The results of experiments show the proposed methodology can achieve better recommendation effectiveness.