Narrow Results

An efficient method to determine consistency in a knowledge-base is described. The traditional generation and evaluation steps are interleaved to compute the differences between two consecutive knowledge-base states. The method employs an optimized update propagation algorithm to improve efficiency. Updates to both the facts and the rule-base are supported by the method.

This paper focuses on an inference methodology based on a belief linguistic rule base (B-LRB) for qualitative decision support. It is termed 'linguistic rule-base' instead of 'fuzzy rule-base' because the use of membership functions associated with the linguistic terms are unnecessary or do not play a key role. The features of B-LRB, the ways to generate a B-LRB, and the inference procedure based on B-LRB are specified, along with an illustrate example applied to evaluate consumer trustworthiness in Internet marketing to show how it works, its applicability and feasibility.