Software engineering utilizes selected knowledge sources from which the fundamental concepts for the software solution can be extracted. The quality of the adopted knowledge sources intrinsically defines the quality of the software solution. This quality of knowledge sources is, on the one hand, determined by its objectivity value, and on the other hand, by its relevance value for the given problem. Since the relevance and objectivity values may change due to newly generated knowledge or evolving requirements, crisp decisions in accepting or rejecting the knowledge sources will result in an inappropriate evaluation. We propose to apply fuzzy reasoning in which knowledge sources are assigned fuzzy linguistic quality values to express the quality degrees. This provides a more precise evaluation and can better cope with the evolution of the knowledge sources and the corresponding software requirements. We describe the validation of our proposal with an experimental case study on the evaluation of domain knowledge for the design of transaction systems.
|Title of host publication||Software Engineering with Computational Intelligence|
|Place of Publication||London|
|Number of pages||20|
|Publication status||Published - 2003|
|Name||Studies in Fuzziness and Soft Computing|
- Fuzzy reasoning
- Domain Analysis