UAP-miner: A real-time recommendation algorithm based on User Access sequences

Recommendation is an application of Web mining. However, most of the current recommendation mechanisms need to generate all association rules before recommendations. This takes lots of time and canpsilat provide real-time recommendations for online users. In this paper, we propose a novel algorithm, called user access pattern mining (UAP-miner), to provide real-time recommendations based user access patterns. UAP-miner uses a data structure named as user access pattern tree, or UAP-tree in short. The user access pattern tree can efficiently store large number of user access information. According to the user access patterns, the UAP-miner scans relevant sub-trees of the user access pattern tree to generate real-time recommendations.