New System for Adaptive Information Retrieval Based on Fuzzy Sets

Search engines are dealing with huge amount of diverse information in many different disciplines. Many retrieval systems attempt to maintain the personalization of the information retrieved through the used search engines. Unfortunately, there are still many problems such as weak concepts representation, inaccurate of retrieved information, and vague retrieval systems. In this research, the fuzzy theory will be used hence; the fuzzy user model can be created. The fuzzy user model contains fuzzy concepts and fuzzy memberships for the acquired user profile which provide an efficient representation for text in addition to adaptive information retrieval. During the main process of the proposed user model acquiring, the document is represented by using the keyword and key phrase extraction. The acquired fuzzy user model is represented in ontological format to be reusable. Such adaptation is achieved by using user model that is based on fuzzy theory to infer fuzzy user modelling aspects. Finally, many advantages are expected from the proposed system such as flexible and reliable retrieval system, overcoming of traditional search engine problems, and reducing of the information retrieval process time. The system evaluation exploits 206 queries, which are generated during five users are searching and browsing the system. Using evaluation metrics; such as precision, recall, fall-out and f-measure, the experiments prove, that the proposed approach improves the accuracy retrieved information rather than other traditional approaches retrieving information.