Title: Mining groups of mobile users

Authors: Dmitry Namiot

Addresses: Faculty of Computational Mathematics and Cybernetics, Lomonosov Moscow State University, GSP-1, Leninskie Gory, Moscow 119991, Russia

Abstract: This paper presents a new approach in data analysis for monitoring of mobile devices. Our model uses passive monitoring of mobile devices based on ideas of network proximity. This monitoring uses network protocol analysis for Wi-Fi and Bluetooth and lets us collect information on mobile visitors. Collected information is a direct analogue for weblog and website usage data, but operates with real visitors (real mobile devices), rather than with abstract requests for web pages. We describe a new approach for processing of these data, which can detect some forms of relationships between mobile users. Our algorithm uses time-based clustering rather than traditional K-means approach. A traditional clustering algorithm (K-means) might find clusters and cluster centres for the given K. For our time stamped events we are not concerned with finding cluster centres. Our algorithm should only assign collected points to clusters as long as the segmentation remains the same.

Keywords: location; mobile monitoring; proximity; Wi-Fi; clusters; mobile users; mobile devices; network proximity; network protocol analysis; Bluetooth; time-based clustering; data processing; data mining.

DOI: 10.1504/IJWMC.2015.073104

International Journal of Wireless and Mobile Computing, 2015 Vol.9 No.3, pp.211 - 217

Available online: 18 Nov 2015 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article