AUTOREB: Automatically Understanding the Review-to-Behavior Fidelity in Android Applications

摘要: Along with the increasing popularity of mobile devices, there exist severe security and privacy concerns for apps. On Google Play, user reviews provide a unique understanding security/privacy issues apps from users' perspective, in fact they are valuable feedbacks users by considering expectations. To best assist end users, this paper, we automatically learn related behaviors inferred analysis on reviews, which call review-to-behavior fidelity. We design system AUTOREB that assesses fidelity employs state-of-the-art machine learning techniques to infer relations between four categories security-related behaviors. Moreover, it uses crowdsourcing approach aggregate review-level app-level. our knowledge, is first work explores review information utilizes semantics predict risky at both crawled real-world dataset 2,614,186 12,783 13,129,783 play, use comprehensively evaluate AUTOREB. The experiment result shows method can app user-review level accuracy as high 94.05%, also app-level aggregating predictions review-level. Our research offers an insight into helps bridge gap perception.