A Fast Spam Filtering Method
Shengyi Jiang · Journal of Chinese Computer Systems · 2013
K Nearest Neighbor algorithm is a kind of classification algorithm for its simple principle and high performance.But because of its high time complexity,it is not applicable to online spam detection.This paper realize a fast spam filtering method which first cluster train emails as initial clusters,and then cluster these initial clusters base on SNN clustering algorithm into final clusters.These final clusters are seemed as classification model which can be incremental updated.In the end,we classify new emails from the classification model.Experiment result on Ling-Spam shows that the method we present in this text not only has higher classification precision but also lower time complexity.