Predicting Human Mobility Using Sina Weibo Check-in Data

Predicting human mobility plays an important role in urban planning and functional zoning. In this paper, we propose a model for predicting the user's next check-in location where indicates that this user is about to arrive at the place based on Sina Weibo check-in data. In our prediction model, the activities are first extracted from the check-in data, then we further classify user's activities using K-Medoids and DBSCAN hybrid algorithm. Finally, we utilize Markov Chain model and our proposed activity detection method to predict the activity category of the user's next check-in location, then further predict the most likely places according to the predicted activity category. The advantage of this model is that we can reduce the prediction space and further improve the accuracy of predicting the next location. Through relevant experimental verification, we not only obtain relatively high activity category prediction accuracy, but also could predict the users' next arrival location conveniently and mobile trends.