Flow characteristic selection algorithm based on dynamic information in deep flow inspection

• Published in: Proceedings of 2011 International Conference on Computer Science and Network Technology Vol. 2; pp. 1216 - 1219
• Main Authors: Guo Lei, Wang Yadi, Yao Qing, Zhu Ke, Yi Peng
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2011
• Subjects: Character recognition; characteristic selection; Classification algorithms; DFI; dynamic information
• ISBN: 1457715864; 9781457715860
• DOI: 10.1109/ICCSNT.2011.6182178

In the technology of deep flow inspection, the recognition and classification of the data flow need using the flow characteristics. The currently characteristic selection algorithm based on the information measurement compute the information entropy of characteristics in the whole sample space, without considering the characteristic selection is a dynamic and changing process, also cannot accurately measure the dependence degree between characteristics in specific selection process. Therefore, this paper puts forward a characteristic selection algorithm based on dynamic information standard, this algorithm takes full account of the changes of information entropy in the characteristic selection process, by removing redundant and useless information, it would achieve the accurate and efficient selection of characteristics. The experimental data shows that, the classification performance of the proposed flow characteristic selection algorithm based on dynamic information is better than the other selection algorithm in the aspect of precision rate and recall rate.