Matching User Accounts Across Social Networks Based on LDA Model

摘要

Identifying users across social networks has received more and more attention in recent years. In this paper we propose a model that combines ATM and DTM. The model uses the ATM model to study the user's potential behavior information and extract user's topic preferences. Then it analyzes the trend of user topic preferences changing with time through the DTM. The purpose of this paper is to improve the accuracy of account identification across social networks. The experimental results show that the model performs well on Chinese text or mixed text in Chinese and English because it obtains higher accuracy and F1 scores.