TEXT UNDERSTANDING SYSTEM

A LARGE VOLUME OF EXISTING INFORMATION IS IN THE FORM OF TEXT INCLUDING MESSAGES, ARTICLES, AND BOOKS. INFORMATION PROCESSING SYSTEMS HAVE NEGLECTED THIS SOURCE OF INFORMATION DUE TO THE LACK OF MECHANISM FOR EXTRACTING RELEVANT INFORMATION OUT OF THOSE TEXTS. INFORMATION RETRIEVAL SYSTEMS ARE MOSTLY EITHER BASED ON STATISTICAL METHODS OR BASED ON KEYWORD IN CONTEXT (KWIC) DOCUMENT RETRIEVAL WHICH BOTH LACK THE NEEDED UNDERSTANDING CAPABILITY. A "CONCEPTBASED" MECHANISM IS PROPOSED HERE FOR EXTRACTING INFORMATION FROM A TEXT-BASE AND REPRESENTING IT IN A FORM USABLE BY AN INFORMATION PROCESSING SYSTEM. IT ALSO PROVIDES A MECHANISM FOR TEXT ABSTRACTION AND CONTROLLED PRESENTATION OF THE INFORMATION THAT IS OF INTEREST TO THE USER. THE TECHNICAL OBJECTIVE OF THE PHASE I ARE TO DEMONSTRATE THE FEASIBILITY OF THE CONCEPT-BASED APPROACH TO TEXT UNDERSTANDING.