Co-clustering Analysis of Mobile Users' Usage Behavior on Apps

Abstract

Nowadays the explosion of smart-phone apps has created a fertile ground to study behavior of mobile users. In this paper, we utilize network footprint data (NFP) which consists of DPI data from ISPs and Crawler data from Web for analyzing the users' behavior. We construct three matrices including (user-app connections matrix, app-time correlations matrix, app-time users matrix) based on the processed NFP data, and apply co-clustering on the matrices. Through analyzing the co-clustering's results in detail, we find many interesting relations among user, app, and time. At last we choose four days' data and analyze the stability of our results.