Interest Mining Model of Micro-blog Users by Using Multi-modal Semantics and Interest Decay Model

Abstract

In order to meet the personalized needs of users and provide better recommendations, how to analyse the user interest accurately has become the focus of research currently. Due to the short content of micro-blog data, the sufficient semantic information it is difficult to get, which leads to the difficulty in accurately mining user interest. Traditional methods mainly use social relations to mine user interest, solving the problem of sparse data to a certain extent. But for users with single social relations, there is a cold start problem, which makes it unable to establish an effective user interest model. In addition, user interest will change over time which results in deviations when using traditional feature extraction methods. In order to solve the problems, we present an interest mining model of micro-blog users by using multi-modal semantics and interest decay model. It builds a connection among semantic relations in multidimensional features. It can solve the data-sparse problem, as well as the cold start problem. The fusion of multiple data semantic represents user interest features more comprehensively. To solve the problem of user interest migration, we propose an interest decay model to assist mining user interest better. In this paper, experiments are carried out on the dataset of 2, 938 user information extracted from micro-blog. The experimental results show that the method proposed in this paper significantly improves the accuracy of user interest mining compared with the existing methods.