专家系统中的语义及其可信度

The Semanteme and Its Confidence Factor in Expert System

针对淮河中游防洪调度决策的实例，用最高最低水位经分段折算，作了语义量化；根据方法库模拟结果来判断某种状态的语义可信度，由半正态分布隶属函数来确定；对各种状态的语义可信度间的转换，采用了模糊聚类的方法。文中提出的方法，已试用于淮河中游防洪专家系统。经多次历史洪水的模拟试验，其可行性和实用性得到了验证。

This paper is aimed at the flood-control expert system for the middle reaches of the Huaihe River. The highest and lowest water levels are adopted to quantize the semantemes. The semi-normal distribution fuzzy function is used to calculate the confidence factors. The calculation is based on the simulated results of methodology-base. The fuzzy classification is used to transfer the confidence factor of one state into another. The suitability and practicality of the above methods have been verified and validated by simulating decision-making of flood-control for many times of historic floods in the middle reaches of the Huaihe River.