Exploiting machine learning techniques for location recognition and prediction with smartphone logs

Abstract

Due to the advancement of mobile computing technology and the various sensors built in the smartphones, context-aware services are proliferating to everyday life. Location-based service (LBS), which provides the appropriate service to smartphone users according to their contexts, is becoming more popular, and the location is one of the most important contexts in LBS. Extracting and recognizing meaningful location and predicting next location are crucial for successful LBS. Many researchers have attempted to recognize and predict locations by various methods, but only few consider the development of real working system considering key tasks of LBS on the mobile platform. In this paper, we propose a location recognition and prediction system in smartphone environment, which consists of recognizing location and predicting destination for users. It recognizes user location by combining k-nearest neighbor and decision trees, and predicts user destination using hidden Markov models. To show the usefulness of the proposed system, we have conducted thorough experiments on real everyday life datasets collected from 10 persons for six months, and confirmed that the proposed system yielded above 90% of average location prediction accuracy.