Research Of Detection Method Of Stealling Privace Application In Android

Since Google company launched the Android smartphone operating system, Android smartphone has quickly and widely popularized. At the same time, Android smartphone has been more and more related to people's daily life with the development of the Mobile Internet. Now, in people's daily life, a lot of private information are stored in Android smartphone. Malware stealing privacy infested on Android system, the security of people's privacy information is under a great threat. Therefore, the research of detection method of stealing privacy malware on Android has important implications.This thesis mainly studies the detection method of stealing privacy malware, the main work is as follows:(1)28 permissions combinations that can get and transfer user's privacy information are concluded through Apriori algorithm and the principle of least privilege. They are called pseudo malicious permissions combinations.(2)A kind of detection method of stealing privacy malware on the basis of malicious degrees is proposed. These pseudo malicious permissions combinations are used as 28 attributes of application, and then, calculate each attribute's weight. Malicious degree is obtained through the calculation of attribute values and the corresponding attribute weights. By comparing application' malicious degree with the malicious threshold we can determine whether the application is a stealing privacy malware. The detection method has been compared with others, which indicates that the detection method is effective.(3)A kind of detection method of stealing privacy malware on the basis of outliers detection algorithm based on the neighbor is proposed. The distance between two objects is calculated by the weighted Jaccard distance formula by using the attribute and the weights of attributes given as above. If an application contains pseudo malicious permissions combination and is classified as outliers relative to benign application data set by the algorithm of the outliers detection based on neighbor, the application will be regarded as a stealing privacy malware. The contrast experiments demonstrate the effectiveness of the above method.(4)A kind of privacy permission control system is proposed, which would control application's privacy permissions fine-grained dynamically.(5)The detection method of stealing privacy malware on the basis of outliers detection algorithm based on the neighbor and the privacy permission control system are embedded within the Android system source code, which enables real-time detection and the control of privacy permissions when an application is installed.