Research On Detection Methods Of Malicious Links For Mobile Telecommunication

With the development of information technology,rapid popularity of 4G networks,and the speedy growth of mobile Internet and mobile intelligent terminals,the malicious links of software present a trend of flood.At the same time,the emergence of pseudobase stations and the serious disclosure of user information,resulting in many users receive malicious messages frequently that carry malicious links,suffering from mobile phone malware links torture.And online banks and other Internet online platform continues to rise,bank phishing links become endless.At present,mobile phone malware links and bank fishing links these two types of malicious links activity is rampant,to people's lives caused by the extremely bad influence,to people's property and even life safety has brought a great threat.In order to guarantee the information security of mobile communication users,this paper proposes a serial joint detection method for Android system malware links,which integrated a black and white list detection module based on domain name,a sensitive keyword detection module and a detection module based on logistic regression algorithm.The sensitive keyword detection module is based on the analysis of the text that carrying the malicious links,to detect whether the text contains sensitive words that induced click;and the logistic regression detection model is constructed based on seven kinds of link features.The above three modules in the same way to progressively work,once draw the conclusion that the termination of the test work is stopped.In order to protect the personal property security of mobile communication users,this paper proposes a method based on bank phishing link and web page sensitive feature detection method,which includes three modules: namely URL filtering module,firstlevel feature detection of the web page module and detection module based on BP neural network algorithm,in which the first-level feature detection of the web page is detected from the web page text features and the web page image features respectively.The construction of BP neural network algorithm is based on the web page text features and the link features.The above three modules in the same way to progressive work,once draw the conclusion that the termination of the test work is stopped.This paper builds Python language simulation environment based on the Linux platform by installing Eclipse software and Py Dev plugin,the serial joint detection method for the malware links of the mobile phone and the phishing page and link features for the phishing links are implemented in the Python language.And the performance of each module was evaluated.The simulation results show that the above method can effectively identify the two types of malicious links in mobile communication,among them,the accuracy rate of mobile malware links detection is 99.5%,and the accuracy rate of bank fishing links detection is 99.0%.