Location prediction based on variable-order Markov model with time feature and user's spatio-temporal rule

초록

Location-based service has been widely used in modern life. It brings a lot of convenience to our lives. Improving the accuracy of location prediction can provide better location- based service. We propose a location prediction method based on the variable-order Markov model with time feature and user's spatio-temporal rule. First, the user's trajectory data needs to be abstracted, and then the useful stay points in the user's trajectory are extracted. The location prediction is performed by scoring each candidate area, and the score is composed of scores in time and space dimensions. Finally, for the possible zero frequency problem, it is solved by mining the spatio-temporal rule of the user. Experiments using the actual data set GeoLife show that the proposed method improves the prediction accuracy. © 2019 ASTES Publishers. All rights reserved.

키워드