Piecing Together App Behavior from Multiple Artifacts: A Case Study Emily Kowalczyk∗, Atif M. Memon∗ and Myra B. Coheny ∗Department of Computer Science, University of Maryland, College Park, MD 20742, USA Email:femily,
[email protected] yDepartment of Computer Science & Engineering, University of Nebraska-Lincoln, Lincoln, NE 68588, USA Email:
[email protected] Abstract—Recent research in mobile software analysis has app is supposed to behave [20], [21]. The marketplaces provide begun to combine information extracted from an app’s source ample space for developers to upload this information, but as code and marketplace webpage to identify correlated variables we have learned, the structure for documentation is loose and and validate an app’s quality properties such as its intended behavior, trust or suspiciousness. Such work typically involves its use open to choice and style [22], [23]. For instance, a analysis of one or two artifacts such as the GUI text, user ratings, developer can write a complete natural language description app description keywords, permission requests, and sensitive API of their app, or they can provide screenshots or videos – or calls. However, these studies make assumptions about how the use a mixture [24]. They might instead, rely on the crowd and various artifacts are populated and used by developers, which provide only a brief description, expecting users to comment may lead to a gap in the resulting analysis. In this paper, we take a step back and perform an in-depth study of 14 popular on the other behavior within their reviews [25]–[27]. Some apps from the Google Play Store.