A User-Oriented Evaluation Mechanism for Mobile Applications Security Based on Behavior Relevance

Abstract

Due to Weak comprehensiveness, insufficient static detection, and failure to consider user’s subjective factors exist in the traditional mobile application security detection. We propose a comprehensive mechanism for application security evaluation based on software behaviour relevancy and user tolerance. This mechanism integrates static analysis, dynamic analysis, and user evaluation, analyzes whether declaration permissions and API calls meet its functional requirements. User-Oriented Mechanism, which incorporates behavioural relevancy, compensates for the coarse-grained static detection and the inefficiency of dynamic monitoring and provides users with a safety reference for mobile applications. The user’s tolerance for the application’s behaviour is introduced in the evaluation phase. The experiments show that the detection framework can detect applications quickly and shallow false positive rate. The evaluation model can reflect users’ expectations, effectively making up for the limitations of existing detection dimensions.