Research on website classification based on user HTTP behavior analysis
Gang Chen · Jisuanji gongcheng yu sheji · 2010
To reduce the computation of website classification and to make the results reflect the user behavior.A set of web pages which have the same URL prefix is regarded as a single object for classification.The features for website classification are extracted from the user HTTP behavior and a scalable decision tree algorithm is used to deal with the large-scale data of the provincial network.Given the website records visited by the users in HERNET,the decision tree model tags the records with news website label,resource sharing website label and communication website label.Compared with the traditional website classification methods,the proposed method needn’t analyze the content of web page and it is suitable for processing large-scale data.