A User-Perceived Freshness Clustering Method to Identify Three Subgroups in Mobile Internet Users

Abstract

The user behavior analysis of the mobile Internet is still to be explored. This increases importance in research and business as the mobile Internet penetration becomes visible. The author performs a preparatory survey to identify the user behavior scenarios during a day. It indicates that there are three distinguished mobile Internet behavior patterns: (a) always active, (b) irregular, (c) prime time use. This gives a novel perspective to understand user interactions in the mobile Internet. Intelligent mobile services require addressing these differences in order to improve user experience. The author proposes a user-perceived freshness method to cluster users with the content access logs in order to utilize this three-subgroup-based view. The revisit ratio in the following month is used to characterize the user clusters and the middle users show a high revisit ratio. The author discusses the characteristics of this middle user segment.