The Application Of Multilayer Document Categorization In User's Preference Mining And Processing

With the rapid development of network technology, a great number ofcorporations are facilitated to enhance their competitiveness by offering user-orientedservices. Users' preference becomes one of the important problems that draw theirattention. By applying data mining on web log files and relevant data, users'preferences can be found and used to support personalized web pages and orientedinformation push. Currently, classification and clustering are two main methods usedin this area.This paper proposed a method other than classification and clustering, i.e.hierarchical dictionary-based multilayer categorization and its application in users'preference mining. This method extracts multilayer features of text documents underthe guidance of a hierarchical dictionary, and stores them in the form of XMLdocuments. With the XML documents, previous results can be fully utilized. Withmultilayer features, suitable layers can be chosen in the categorization process toensure a satisfactory categorization result. The result of this method, i.e. user'spreferences and user groups' preferences, can also contain multilayer features, andthey are also stored in the form of XML documents.Hierarchical dictionary-based multilayer categorization can fully utilize theprevious results, and can generate better categorization result. It can provide moreefficient decision support in offering personalized web services.