Research on URL Classification with DPI Data Mining and Related Technology

Abstract

In order to achieve the precise localization of internet customers' preference, the crucial point is to catch the classified mapping of URL system through digging deep into the data of DPI users' online behavior data. Based on summarizing the process of DPI data auto classification, the key technique was investigated which involved in Web information extraction, Chinese word segmentation, feature selection, text classification and so on. It may make solid technical foundations for the customer insight with DPI data.

Keywords

• deep packet inspection;Chinese word segmentation;feature selection;text classification