| In most knowledge based systems, the knowledge is primarily heuristic in nature. Such heuristic knowledge is sometimes referred to as "shallow" knowledge because it does not contain much information about the causal principles underlying it. In order to overcome the problems associated with such heuristic knowledge, like sharp degradation of performance outside the narrow domain, it has been suggested that knowledge based systems should be "model based". Qualitative reasoning provides us with explicit methodologies for developing an underlying domain model. Because of that, knowledge based systems developed on the principles of qualitative reasoning should be able to avoid the problems associated with heuristic knowledge based systems like rule based systems.;However, before qualitative reasoning is used on a large scale to develop knowledge based systems, the effectiveness of using qualitative reasoning has to be known. The primary objective of this research is to measure the effectiveness of qualitative reasoning against a rule based system. To measure this effectiveness a qualitative reasoning system as well as a rule based system has been developed for forecasting the net income of an organization. The system with better forecasting ability is considered to be more effective for forecasting.;While the primary purpose of this research is to study the effectiveness of qualitative reasoning as measured by its forecasting ability, a secondary purpose is to study how the two systems differ in structure and reasoning. |
