A Path Summarization and Prediction Method Based on Meaningful Locations

Abstract

In this paper, we propose a path prediction approach using behavioral data of user that it contains the meaningful locations extracting and predicting method. The proposed method has a difference to the previous methods that is considering the interaction data for defining the meaningful location and predicting the future paths. Using these interaction and path data, the proposed method calculates the proximities of adjacency people. For extracting the meaningful locations, we consider the calculated proximity of people around user and stay time based on path data. And we simplify the paths using these extracted meaningful locations. Finally, in prediction step, the method predicts the destination using the simplified paths, and finds detail path from current location to destination using modified Dynamic Time Warping (DTW) algorithm. For verifying the usability of proposed method, first, we analyze the effect of people around the user for predicting the future paths of user. We verify the effectiveness of the proposed method by comparing the prediction accuracies of each method.