Intraday-scale long interval method of classifying intramonth-scale revisiting mobile users

Abstract

Penetration of the mobile Internet has increased its visibility worldwide. This enables analysis of detailed time-dimensional user behavior data. It also increases the industry need to identify and retain mobile users with strong loyalty to a particular mobile Web site. The author proposes an intramonth-scale revisit classification method for identifying intramonth-scale, revisiting mobile users. The author performs a case study and the result shows that the proposed method shows 87 % classifier accuracy. The author discusses a trade-off between classifier accuracy and a true positive ratio. © 2008 International Federation for Information Processing.