Generating and Applying Rules for Web Documents Retrieval

Web documents retrieval is very challenging due to the huge amount of documents available and difficulty to interpret these documents. Both effectiveness and efficency of retrieval are important. This paper presents some approaches from soft computing to improve effectiveness of web documents retrieval. These approaches give a more accurate and reasonable representation of terms provided by the user, present how to match terms and documents with fuzzy logic techniques, and show how to reduce the number of matched documents and necessary terms. A possible architecture for a neuro-fuzzy system to match terms and documents is sketched. This paper also discusses how to form linguistic rules by polling a panel of experts.