Predicting Locations of Mobile Users Based on Behavior Semantic Mining

摘要

Predicting movements of mobile users has become increasingly popular because of the ease of trajectory data collecting nowadays. However, as most of these prediction techniques need geographic pattern matching of users' trajectory data, it is possible that the techniques cannot work in a place where the user has never been before. In this paper, we propose an approach based on transportation mode and behavior semantic features to predict the next location of the users' movement. First, we identify the users' transportation mode to get sequential data of the users' motion mode. Then, we get the semantic meaning as behavior semantic features from the places where users have stopped and visited for a while. We determine the relationship between the transportation mode and behavior semantic features to predict the next location based on the Hidden Markov model. We use real world data for our experiment to demonstrate the effectiveness of our approach.