A deep learning-based android malware detection method

Abstract

The invention relates to a deep learning-based Android malware detection method, which belongs to the technical field of computer and information science. The invention first extracts the features of the Android application software, and then extracts the relevant security features through operations such as decompressing and decompiling the Android application files. The extracted features include three aspects: file structure features, security experience features, and N-gram statistical features composed of Dalvik instruction sets. Then the extracted features are numerically processed to construct feature vectors. Finally, a DNN (Deep Neural Network) model is constructed based on the above extracted relevant features. Classify and identify new Android software through the constructed model. This method combines the analysis of the instruction set, which has the effect of resisting malware confusion. At the same time, the malware detection based on the deep model can enhance feature learning, can well express the rich internal information of big data, and is easier to adapt to the evolving malware. .

Description

technical field [0001] The invention relates to a deep learning-based Android malware detection method, which belongs to the technical field of computer and information science. Background technique [0002] With the continuous development of the mobile Internet, smart terminals have become an important part of everyone's life. Android, as the most widely used mobile operating system, has led to the proliferation of malware due to its open and flexible ecological environment. How to effectively detect Android malware is a research topic with great value. The current mainstream Android malicious code detection methods are roughly divided into static detection methods and dynamic detection methods. [0003] 1. Dynamic detection method [0004] The so-called dynamic detection and analysis refers to the method of extracting features for detection and analysis after the detected program is running. The dynamic detection method is mainly to analyze the software by running the ...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to solve the problems of low detection accuracy, limited detection scope, and difficulty in adapting to emerging software in conventional Android malware detection methods, and propose a malware detection method based on deep learning

Images