Accession Number : ADA279897


Title :   Case-Based Reasoning in Mixed Paradigm Settings and with Learning


Descriptive Note : Final rept. 15 Sep 1990-28 Feb 1994


Corporate Author : MASSACHUSETTS UNIV AMHERST DEPT OF COMPUTER AND INFORMATION SCIENCE


Personal Author(s) : Rissland, Edwina L


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/a279897.pdf


Report Date : 30 Apr 1994


Pagination or Media Count : 14


Abstract : In this project we investigated: (1) CBR in mixed paradigm settings, in particular in a blackboard-based system, called FRANK, that generated various types of explanations and arguments where supporting tasks, such as case-and rule-based reasoning, were dynamically configured to reflect the user's intended purposes for the report: (2) pure CBR, particularly issues concerning the use of multiple indices and types of case representations, in a system called BankXX. that used classic heuristic best-first search to retrieve information needed for case-based argument: and (3) the application of machine learning techniques to core issues in CBR, such as the problems of learning indices and prototype cased and estimating concept theory drift.


Descriptors :   *HEURISTIC METHODS , *TAXONOMY , REASONING , THEORY , PROTOTYPES , LEARNING , DRIFT , INFORMATION RETRIEVAL


Subject Categories : Computer Programming and Software


Distribution Statement : APPROVED FOR PUBLIC RELEASE