Home // MOBILITY 2014, The Fourth International Conference on Mobile Services, Resources, and Users // View article

Predicting Destinations with Smartphone Log using Trajectory-based HMMs

Authors:
Sun-You Kim
Sung-Bae Cho

Keywords: destination prediction; forecasting; hidden Markov model; location-based service

Abstract:
With the spread of smartphones, it is easy to obtain sensor data from users, and location-based service (LBS) becomes the most common service in mobile industry. Predicting the user's destination can lead to a variety of services in mobile devices. In addition to the user’s final destination, the locations during movement are also important in LBS. In this paper, we propose a destination prediction method based on hidden Markov models for representing the paths using sensor data from smartphone and identifies the destination and intermediate locations in future moving with visiting probabilities. In order to demonstrate the usefulness of the proposed method, we compare it with Dynamic Time Warping (DTW), a method of template matching. Experiments with the data collected by 10 college students for five months confirm that the proposed method results in 12.67 times faster and 2.88 times more accurate than the DTW.

Pages: 6 to 11

Copyright: Copyright (c) IARIA, 2014

Publication date: July 20, 2014

Published in: conference

ISSN: 2308-3468

ISBN: 978-1-61208-366-7

Location: Paris, France

Dates: from July 20, 2014 to July 24, 2014