User-selectable interaction and privacy features in mobile app recommendation (MAR).

The research in the mobile app recommendation system (MARS) is steadily increasing due to the involvement of big data and social information. Few existing systems explore the possibility of social interaction information and generate accurate recommendations. However, the role of social network-based app recommendations and associated privacy issues are not fully explored because social information from the mobile phone is difficult to access. To refine and construct such a system, it is imperative to understand users' opinions on the type of social interaction information to be included and concerning privacy issues. To achieve this, we conducted a questionnaire survey. We collected data from 1,006 participants, and tested our hypothesis using PLS, Blindfolding, Bootstrapping, and Machine Learning approaches. After evaluation, we consolidate a final feature list related to social interaction and privacy aspects. For practical evaluation, we implemented a prototype using a few social features identified in the questionnaire phase. Our results indicate an increase in click rate up to 0.66. [ABSTRACT FROM AUTHOR]