Immune-Based Dynamic Updating Algorithm
This chapter presents the immune based dynamic updating algorithm using support vector machine (SVM) and artificial immune system (AIS) as the basis for uninterrupted detection approaches for spam filtering . The first three sections introduce the background knowledge, including exceeding margin update (EM-update) principle and sliding window, which are two of the core strategies employed in the strategies. After description of the primary and secondary responses in BIS, overview of the dynamic updating algorithm is given by mimicking some of the mechanisms or functions in BIS. We then give the whole implementation process and related factors of the dynamic updating algorithm in detail, including message representation, dimension reduction, initialization of the window, classification criterion, update of the classifier, and purge of out-of-date knowledge. Finally, experimental validation and discussion are shown.