摘要
针对专家系统多证据规则中各证据对结论的重要程度不同,且同一证据在不同的时期对结论的重要程度也不同的特点,提出了为证据引入动态加权因子的不确定推理方法。该方法使证据的权值在不同时期能够根据证据发生的次数适时改变,使证据的权值分布更合理,使不确定推理的结论真值更准确;还提出了动态规则强度和动态规则阈值的计算方法,并通过算法实例验证了该方法的有效性和实用性。
Each of evidences has a different importance on the conclusion in the multiple evidence rules of expert system,and the same evidence has different importance in different periods.Taking into account the above characteristics,a dynamic weighted factor method for evidence in uncertainty reasoning is presented in this paper.This method allows the weight of evidence to change at different time according to the number of evidence occurrence,so that the distribution of the evidence weight is more reasonable and the truth value of uncertainty reasoning is more accurate.In addition,we also propose a method of calculating dynamic rule certainty and dynamic threshold value.The availability and practicality of this new method has been verified by an algorithm example.
出处
《机械工程与自动化》
2012年第2期204-206,共3页
Mechanical Engineering & Automation
关键词
动态加权
不确定推理
动态规则强度
动态阈值
dynamic weighted
uncertainty reasoning
dynamic rule certainty
dynamic threshold value
