Android platform malicious application detection method and device

Abstract

The invention provides an Android platform malicious application detection method and device. The method comprises the steps that a FlowDroid tool is called, and static data stream features of a to-be-detected Android application are extracted; an SUSI technology is utilized to process the static data stream features of the to-be-detected Android application, and feature vectors of a data stream of the to-be-detected Android application are generated; and the generated feature vectors of the data stream of the to-be-detected Android application are input into a deep belief network detection model trained in advance, and a detection result about whether the to-be-detected Android application is a malicious application is obtained. Through the Android platform malicious application detection method and device, an Android platform malicious application can be accurately detected, the path cover problem existing in dynamic stain tracking is avoided, the two major challenges that accurate modeling needs to be performed on an application running process and target components in communication among components need to be accurately acquired according to a static data stream analysis technology are overcome, accurate and comprehensive extraction of the Android application sensitive data stream is realized, and meanwhile the limitation existing when a detection model is constructed through a traditional shallow machine learning algorithm is overcome.

Description

technical field [0001] The invention relates to the technical fields of mobile security and machine learning, in particular to a method and device for detecting malicious applications on an Android platform. Background technique [0002] In the field of mobile smart terminals, there are a large number of malicious software, which can covertly obtain the private data stored on the device by the user without the user's awareness, and send it to the attacker's mailbox or server, which is harmful to the user's financial security and privacy. Security is a big concern. With the popularity of smart terminals on the Android platform, privacy theft attacks and malicious application detection technologies on Android smart terminals have attracted more and more attention. [0003] At present, the existing data flow analysis technologies for detecting malicious applications on the Android platform mainly include: dynamic taint tracking and static data flow analysis technologies. Dyna...

AI Technical Summary

Problems solved by technology

At the same time, the traditional machine learning algorithm has a shallow model structure, which has a certain impact on the final detection results

[0007] In view of this, how to provide a method and device for detecting malicious applications on the Android platform, so as to avoid the path coverage problem existing in dynamic taint tracking, overcome the need for accurate modeling of the application operation process and the need to accurately obtain communication information between components in static data flow analysis technology The two major challenges of the target component are to achieve accurate and comprehensive extraction of Android application sensitive data streams, and at the same time overcome the limitations of traditional shallow machine learning algorithms in building detection models, and to achieve accurate detection of malicious applications on the Android platform has become a current need. Technical issues resolved

Images