Design And Implementation Of Spam SMS Filter System Based On MTK Platform MobilePhone

In China, mobile telephone users take the SMS as the common way to contact friends for its convenience and low price. Recent years, with the development of mobile communication technique and the increase of mobile users, SMS service is also well received, but the spam SMS spreads quickly at the same time, which is a severe disturbance to people’s daily life. Therefore, the spam SMS becomes a hot issue. For the communication operators, to filter spam SMS without any mistakes is a hard work and the data needed to be handled with is too much while the real-time requirement of SMS is as high as possible, so to filter spam SMS is focused on clients. Nowadays, mobile phones based on MTK platform still play an important role in domestic markets, thus we should study methods of filtering SMS of MTK phones since it lack of applications of it compared to intelligent phones.In this thesis, a system of spam SMS filtering based on blacklist and keywords is completely designed and studied, which brings in a highly efficient matching algorithm on shortcomings of general systems and proposes an effective way of updating the keywords library to allow users take advantages of blacklist and keywords to filter spam SMS. Meanwhile, using the decision tree algorithm for the system can achieve a better performance of filtering.The following tasks are accomplished in this thesis:Firstly, the thesis studies associated techniques of SMS filtering system, including keyword-filtering algorithm, text segmentation, feature extraction, text classification and so on, and mainly describe keyword-searching algorithm and application based on decision tree algorithm of the system.Secondly, it designs the system that combines the blacklist and keyword-filtering by connecting the software with hardware of MTK phones, and analyzes key techniques and problems of original system and use WM(Wu-Manber) multi-mode matching algorithm to matching keywords which improves the efficiency of searching. In order to improve the system, the thesis comes up with the way of log feedback to analyze the filtered spam SMS, and then extracts keywords to update the keywords library.Finally, it applies the decision tree algorithm to the system, and draws seven features as the input parameters to acquire decision tree and generate the corresponding decision rules. This thesis also tests and assets the system, the result indicates a good performance of it.The system has been applied to several MTK platform mobile phones, all the test data shows that it stable and reliable to operate, and the feedback is good, moreover, various aspects of performances satisfy the requirements of the issue.