An expert system for searching in full-text

摘要

This project applies expert system technology to the task of searching online full-text documents. We are developing an intelligent search intermediary to help end-users locate relevant passages in large full-text databases. Our expert system automatically reformulates contextual Boolean queries to improve search results and presents retrieved passages in decreasing order of estimated relevance. It differs from other intelligent database functions in two ways: it works with semantically unprocessed text and the expert system contains a knowledge base of search strategies independent of any particular content domain. The goals for our current project are to demonstrate the feasibility of the approach and to evaluate the effectiveness of the system through a controlled experiment. While the work we report here has limited objectives, the system and techniques are general and can be extended to large, real-world databases.