The Similarity Calculation of E-commerce User Behaviors with Petri Net

Abstract

The similarity of user behaviors is conductive to classify users, cluster different groups, and generate the behavior model. It is of great significance to achieve a finegrained analysis of user behaviors. In this paper, a novel method of computing the similarity of e-commerce user behaviors is presented based on Petri nets. This method involves the following aspects. We first use Petri nets to model user behaviors. Then, we calculate the similarities of behavioral sequences and behavioral data, respectively. Finally, we get the results by comparing the similarities of users. Our experiment presents that this method can reflect data features and behavioral sequences by places and transitions of Petri nets. Above all, it shows a higher accuracy of the similarity calculation.