An intelligent technique for extracting subjects from user profile using ODP ontology-driven reasoning

作     者：Minazadeh, Mina Badie, Kambiz Pedram, Mir Mohsen 

作者机构：Department of Computer Engineering Islamic Azad University Science and Research Branch Tehran Iran Information Technology Research Faculty Telecommunication Research Center Tehran Iran Computer Engineering Department Faculty of Engineering Tarbiat Moallem University Karaj/Tehran Iran 

出 版 物：《Research Journal of Applied Sciences, Engineering and Technology》 (Res. J. Appl. Sci. Eng. Technol.)

年 卷 期：2012年第4卷第15期

页      面：2334-2342页

核心收录：

学科分类：1205[管理学-图书情报与档案管理] 12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 

主　　题：Ontology 

摘      要：Nowadays, the amount of available information, especially on the Web, is increasing. In this field, the role of user modeling and personalized information access is obviously vital. The traditional techniques like BOW (Bags of words) limit recommendations to the words which have been stored in the profile. In other words, the news items, which semantically relate to the users interests, can t be recognized and recommended to the users. Besides, BOW technique suffers from the curse of dimensionality, thus computational burden reduction is an essential task to efficiently handle a large number of terms in practical applications. This study focuses on the problem of choosing a representation of documents that can be suitable to induce concept-based user profiles as well as to support a content-based retrieval process. In this study, a new approach has been proposed to construct a ranked semantic user profile through extracting the related subjects. The new items can be recommended through collecting information from the user s selections, based on existing domain ontology ODP. The efficiency of the proposed technique has been shown by embedding it into an intelligent aggregator, RSS (RSS is acronym of Really Simple Syndication) feed reader, which has been trained and evaluated by different and heterogeneous users. The results in experimental session show that the incoming news item which semantically relates to the profile gets highly recommended to the user despite its excluding of common words in the profile. © Maxwell Scientific Organization, 2012.