In this paper we propose a mass market solution on mobile phones to discover a user's significant places solely from observed cell IDs. It does not require either cell-ID-to-physical-location mapping or the capability of obtaining multiple cell IDs on the phone simultaneously, and is able to run on virtually any mobile phone today. Our solution is centered around a cell ID clustering algorithm based on temporal correlations. It is able to prevent over-clustering and handles missing data well. We evaluate the solution with real-life data that the author has collected over a period of eight weeks. Results show that we are able to discover not only places of utter importance, but also certain less frequently recurring places and one-time travel destinations that bear significance in one's life. © 2009 Springer Berlin Heidelberg.
CITATION STYLE
Yang, G. (2009). Discovering significant places from mobile phones - A mass market solution. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5801 LNCS, pp. 34–49). https://doi.org/10.1007/978-3-642-04385-7_3
Mendeley helps you to discover research relevant for your work.