A Long Interval Method to Identify Regular Monthly Mobile Internet Users

The Internet access from mobile handsets is extending its coverage. It increases the importance to analyze mobile Internet use in research communities and industries. To cope with the increased demands on mobile Internet user analysis, the author describes a regularity-oriented service model. In order to capture user regularity, the author performs an exploratory clickstream analysis to identify users with high loyalty. Based on the author's previous experience on the relationship between middle-range visit intervals and revisit ratio in the next month, the author tries to identify hour-scale intervals linked to the high-loyalty-users. The author obtains the result that the users with intervals with 480 minutes and 720 minutes show the high revisit ratio. The author analyzes key time zones for high- loyalty users. This method is applicable to a wide range of applications with clickstream logs. Using the long-interval characteristics, the author discusses a double peak model of sub-day-scale activity behaviors in the case study.