Mobile User Drop Out Ratio Observation with Time zone and Day-of-week Regularity Analysis

The mobile Internet fills the gap between 24-hour user behavior and Internet clickstreams. The mobile-Internet specific methodology to capture the user behavior characteristics is on the hot research agenda. The long-term stability of the mobile user behavior characteristics is still unexplored. The author proposes a method to identify the user behavior patterns using transition pattern clustering over a long span of time. The author gives an exploratory analysis to find best-fit threshold value to identify regular user behaviors with a 3-month transition patterns. From this observation, the day-of-week behavior gives better indication compared to that of time zone-based one.