Abstract: The research of monitoring and filtering of the files transporting through internet is getting hotter and hotter now．The traditional algorithm based on string-matched is not able to meet the need of the huge increase of information．Although SVM model can surely improve the efficiency of the classification,the problem that SVM’s too large dimension will affect the speed of examine still exists．It also causes a waste of storage space and compute ability．One algorithm was raised by first reducing the dimension by some specific algorithm before classification．The analysis result shows that after the improvement,we can get a more accurate result.

Key words: Feature reduce,One-class SVM,Classification