What is the Knowledge database Viewer?

The Knowledge database Viewer is a program which allows you to search through multiple knowledge databases of information. This process of searching through existing cases. or databases of information, is called Case-based Reasoning (CBR).

The Knowledge database Viewer (KBVIEWER) is designed to work in a standalone mode, as well as in integrated mode. In integrated mode, other applications can pass Search Requests to the KBViewer, which will then search for the results and then display, or report on them.

The KBViewer subsystem works with Windows, Windows for Workgroups, and Windows NT.

What is Case-Based Reasoning?

Case-based reasoning (CBR) is a new approach to decision automation. CBR can dramatically reduce the costs of developing, maintaining, and running an expert system. Traditional expert systems attempt to model human problem-solving expertise as a purely deductive process. They develop a unique new solution to each problem. However, there is evidence that human experts do not work in this manner. Instead, a human will remember and adapt solutions from previous problems. Given this observation, a number of artificial intelligence researcher have focused on solving problems by first retrieving previous solutions rather then generating new ones. This is case-based reasoning (CBR), the technique of comparing a current case to a database of cases with known solutions.

Previous solutions are stored as "cases" in a knowledge database (or case-database or simply casebase). The case database can be used for multiple purposes, including training and decision support as well as decision automation. When a search through the case database fails to locate a similar case (or provide a "match"), the search itself becomes the basis for a new case, in effect learning from the experience. Because of this, a CBR system can stay current in a changing environment. Because of the simple representation of Cases, Questions, and Actions, little expertise is required to maintain a CBR system. The Case Administrator does not need to be a programmer (as most expert systems would require) !

The advantages of the CBR approach are evident in terms of dealing with PROBLEMS. Resolving PROBLEMS through the searching in a case database allows the very process of tracking problems to become one of providing more knowledge in the hands of the problem solver. In a case-cased reasoning system for help desk automation, knowledge about problems and their resolutions are stored as cases. The cases are indexed by features which uniquely describe each case. When presented with a problem description, the knowledge database viewer will retrieve the closest matching cases, and a list of questions to be asked to help confirm or eliminate some of the cases. Because CBR uses advanced reasoning, the process of searching can identify partial matches (as opposed to exact matches in a traditional database system), the CBR system can work with little or incomplete, or even incorrect information. The solution to the problem will be based on the solution(s) of the cases that are most similar to the problem.