One-path Relaxed Realtime Constraint Mobile User Classification Method in Mobile Clickstreams

Abstract

The mobile Internet has become increasingly visible in everyday life. As mobile Internet penetration leverages content business opportunities, it is crucial to identify methodologies to fit mobile-specific demands. Regularity is one of the important measures to capture easy-come, easy-go mobile users. It is known that users with multiple visits per day with a long interval in between have a higher likelihood of revisiting in the following month than other users. The author proposes a 3+1bit method to incorporate this empirical law in order to cope with the two major mobile restrictions: distributed server environments and large data streams. The proposed method can be performed in a one-path manner with 32-bit word boundary-awareness for memory compaction. Experimental results show that the method is promising for identifying revisiting users under mobile-specific constraints.