Web Usage Prediction and Recommendation Based on Web Session Graph Embedded Analysis
Shuning Huang, Kaoru Ota, Mianxiong Dong · 2020
Web usage prediction and recommendation have validated its importance through the enormous economic benefits it brings in areas such as e-commerce. It also plays an essential role in the low latency services needed by the new generation network services. As a usage record of network user interaction with the network, the network session logs contains the user's preferences and behaviour rules. In this paper, we propose a new web usage prediction and recommendation model by graph learning the potential laws of web session logs and achieving recommendations by anticipating the future behaviour exploiting prediction cluster. The model values multiple sets of information recorded in the session logs to master its consistency and complementarity and enhance the accuracy of the task. Experimental results on real session dataset show that the proposed model is reliable and efficient.