Mining Sequential Application Usage Patterns with Time Constraint of Smartphone Users

Abstract

Smartphones are mobile phones with extended functionality and have brought changes to our daily lives. The main feature of smartphones is that the developers can create applications that users can use to personalize their smartphones. As both the number of smartphone users and the number of available smartphone applications increase, it is practically important to study how users use the applications on their smartphones. Thus, we apply a sequential pattern mining technique to a set of smartphone logs collected from several users over several months. This paper presents a data preprocessing process that transforms the logs to sessions. It also presents an algorithm designed to mine sequential patterns with constraint on the maximum time interval between applications used in a session. It further presents interesting patterns generated by using different time intervals. This paper contributes to behavior analysis of smartphone users by providing an approach used to discover patterns from smartphone logs and by providing a discussion about the discovered patterns. What presented in this paper could be beneficial to the developers of smartphone applications.