Improved ReliefF algorithm applied in detecting image spam
Wan Ming-cheng · Jisuanji yingyong yanjiu · 2009
The spread of image spam brought much a pernicious influence on society.The application of image spam filters holds back the spread of image spam.At the depletion of time and the accuracy,they can't satisfy both sides.This paper proposed an ReliefF feature selection algorithm based on feature redundancy(R-ReliefF algorithm),getting the features of iamge spams firstly,and discreted the features according to property of the featuresthe,then applied it on the optimization of the set of image spam features discreted.According to contradistinction,the optimization subset of features both decreased the depletion of time and increased the accuracy on detecting the image spam.