Research On Android Malware Detection Based On Machine Learning

The birth and development of 3G,4G and 5G are extending the bandwidth of mobile internet continuously and Mobile Internet Era has changed the way of human life.Meanwhile,smart mobile devices are also in rapid development and spread.The market of smartphone and tablet computer has reached 1.43 billion in 2015,while 81.2% of these devices are running Android OS.Related parties are paying close attention to security issues of Android because it is the major platform of smart mobile devices.The reaserch and application of effective Android malware detecting technology become crucial as well.In this context,this paper analyzes the status of Android malware detection,the advantages and disadvantages of various detection methods,and farther proposes a scheme for dynamic detection of Android malware based on machine learning.In order to verify the effectiveness of the scheme,this paper collected the corresponding experimental sample databases and conducted the actual detecting experiments.First of all,this paper introduces the basic technical background,including Android system and its security mechanism,situation of malwares,malicious actions and detecting techs.Then,the dynamic detection technology and the machine learning method are introduced.On this basis,this paper presents a dynamic detection scheme based on machine learning and describes the details.This scheme uses both Strace and DroidBox dynamic features for machine learning classification.The next part describes the experimental settings,the experimental verification results and analysis.Android malwares are still in a fast evolution,bring new kinds of malicious actions continually.The malware detection scheme based on machine learning proposed in this paper,proved to be an effective solution to this,and it is worthy of further study.