The mobile Internet enables 24-hour always-on comput- ing with increased demands on time-based analysis. From an assumption that the mobile users are split into three dif- ferent clusters, the author proposes a method to identify the three different user clusters. The author evaluates the user segments in a subscription-based commercial service using the revisiting ratio in the next month.
Citation:
Toshihiko Yamakami, "Exploratory Day-Scale Behavior Assumption-Based User Clustering with the Mobile Clickstream," pdcat, pp.169-170, Eighth International Conference on Parallel and Distributed Computing, Applications and Technologies (PDCAT 2007), 2007