A Structured Testing Methodology for Knowledge-Based Systems

Abstract

In recent years, knowledge-based software technology has proven itself to be a valuable tool for solving hitherto intractable problems. Developers of knowledge-based systems must ensure that the system will give its users accurate advice or correct solutions to their problems. Thus, knowledgebased systems must be debugged and validated just like any other piece of software. It has been found that one of the most important problems in developing knowledge-based systems is the lack of methods to verify and validate its KB. The aim of this article is to define a methodology and its supporting tool set that are used together in order to completely test knowledge-based systems. The suggested testing methodology couples different verification and validation activities that are collectively valuable in raising the level of system correctness 1.Introduction Knowledge-based systems (KBSs) can be defined to be "A computerized system that uses knowledge about some domain to arrive at a solution to a problem from that domain. This solution is essentially the same as that concluded by a person knowledgeable about the domain of a problem when confronted with the same problem" [Avelino & Douglas, 1993]. One fundamental characteristic of all KBSs is the clear and clean separation between the knowledge that the system is using and the program that utilize it for problem solving. The two components that compose the intelligent program are therefore, a knowledge base (KB) and an inference engine (IE). Since the inference engine is algorithmic software i.e., conventional software, software engineering testing techniques can be applied on it. Therefore, in the field of AI testing for KBSs is limited to testing of KB [Vicat, Brezillon & Nottola, 1995]. In fact, the quality of a KBSs is often adequate to the quality of the knowledge stored in the KB [Smith & Kandel, 1993]. Our suggested testing methodology focus on improving the correctness related aspects of the KB that go towards the creation of high quality KBSs.

A Development-Oriented IS Evaluation Approach: Case Demonstration for DSS

“…This finding also applies to specialised DSS development such as ES and KBDSS. As described earlier, conventional evaluation approaches for specialised DSS development are based on the V & V (verification and validation) approach (Mahaman,Passam, Sideridis and Yialouris 2003;Mansingh, Reichgelt and Bryson, 2007;Juristo and Morant, 1998;El-korany, Rafae,Baraka and Eid, 2000). Verification is described as "building the system right", that is, a system appropriately implements its specification.…”