Mining Android Apps to Predict Market Ratings

Eric Shaw; Alex Shaw; David Umphress

6th International Conference on Mobile Computing, Applications and Services

Research Article

Mining Android Apps to Predict Market Ratings

Download674 downloads

Cite: BibTeX Plain Text

@INPROCEEDINGS{10.4108/icst.mobicase.2014.257773,
    author={Eric Shaw and Alex Shaw and David Umphress},
    title={Mining Android Apps to Predict Market Ratings},
    proceedings={6th International Conference on Mobile Computing, Applications and Services},
    publisher={IEEE},
    proceedings_a={MOBICASE},
    year={2014},
    month={11},
    keywords={android market rating data mining quality},
    doi={10.4108/icst.mobicase.2014.257773}
}

Eric Shaw
Alex Shaw
David Umphress
Year: 2014
Mining Android Apps to Predict Market Ratings
MOBICASE
IEEE
DOI: 10.4108/icst.mobicase.2014.257773

Eric Shaw¹^,*, Alex Shaw¹, David Umphress¹

1: Auburn University

*Contact email: ess0006@auburn.edu

Abstract

Market rating systems give Android users the opportunity to provide feedback on an application (app). Developers aspire for the highest ratings possible, as they reflect upon user perceptions of their apps. However, no mechanism exists to predict in any way the market rating of an app before publication. We downloaded and reverse-engineered 10,740 apps from the Slide Me market, and analyzed them using quality related metrics. We compared the results of the 1,000 highest rated apps against the lowest rated 1,000. Our results show that traditional white box quality metrics do little to distinguish the groups, while certain Android specific user-perspective metrics are useful in prediction.

Keywords: android market rating data mining quality

Published: 2014-11-28
Publisher: IEEE

: http://dx.doi.org/10.4108/icst.mobicase.2014.257773