Characterizing and predicting the cross-app behavior in mobile search.

Purpose: This paper aims to explore the users' cross-app behavior characteristics in mobile search and to predict users' cross-app behavior using multi-dimensional information. Design/methodology/approach: This paper presents a longitudinal user experiment in 15 days. This paper recruited 30 participants and collected their mobile phone log data in the whole experiment. The structured diary method was also used to collect contextual information in mobile search. Findings: This study focused on the users' cross-app behavior in mobile search and described cross-app behavior's basic characteristics. Usage of communication app and tool apps could trigger more cross-app behavior in mobile search. The method of cross-app behavior prediction in the mobile search was proposed. Collecting users' more contextual information, such as search tasks, search motivation and other environmental information, can effectively improve the prediction accuracy of cross-app behavior in mobile search. Practical implications: The future research on cross-app behavior prediction should focus on context information in mobile search. Better prediction of cross-app behavior can reduce the users' interaction burden. Originality/value: This paper contributes to research into cross-app behavior, especially in the mobile search research domain. [ABSTRACT FROM AUTHOR]