Research On Spam-filtering Technique Based On Immunity

With the increasing application of E-mail, how to resolve the serious problem of spam effectively has become the industry's urgent need and critical mission, which will improve people's working efficiency, reduce network security risks and reduce people's losses. Artificial immunology is a new research field in computing intelligence. Researching and developing the system based on the immune system will improve the capabilities of immune technique in a large extent, and will meet the development of modern computer systems and artificial immune systems.Based on above analysis, this article begins with the basic theory of the e-mail, analyses the advantages and disadvantages of several anti-spam technologies in detail which are used commonly today, and studies artificial immune system and its core algorithms deeply, such as negative selection algorithm and clone algorithm. On this basis, this dissertation presents a spam filtering technique based on artificial immune, which will enhance the performance of spam filtering technique. The artificial immune algorithm is the core of the new spam filtering technique, which makes new technique have better characteristics, such as adaptive, self-study and robustness, and so on.The emphasis of the research subject is immune theory in spam filtering applications. In view of the characteristics of E-mail filtering system, the dissertation makes the following job:firstly, the dissertation improves the string matching mechanism. Secondly and thirdly, the dissertation improves the algorithm of generating new antibodies and clone algorithm to make them be more in line with the spam filtering system. Finally, the dissertation analysises the system's dynamic and self-adaptive, and analysises the impact of characteristic quantity, threshold and corpus pretreatment levels.Finally, the dissertation uses the largest positive matching method to implement segmentation technique, uses the immune model to classify E-mail. Through the systematical testing, it shows that using the immune technique in spam filtering is one of the effective ways.