CAJ | 학술논문

为克服传统贝叶斯网络进行态势估计时，在主观性环境下条件概率难以获取和指定条件概率表的工作量随贝叶斯网络中节点数量的增长呈指数增长的问题，提出了用影响网络对态势估计进行建模的方法；基于影响网络的态势估计模型，在继承传统贝叶斯网络结构的基础上，利用促进与抑制两种因果影响关系使建模更为直观、易懂，更好地表现证据节点对其子节点的影响关系；利用CAST逻辑使条件概率表的获取更为简单，解决了条件概率随节点增长而指数增长的问题，并能很好地表现专家知识以及为条件概率表的获取时提供了一种新的解决方法。
위극복전통패협사망락진행태세고계시，재주관성배경하조건개솔난이획취화지정조건개솔표적공작량수패협사망락중절점수량적증장정지수증장적문제，제출료용영향망락대태세고계진행건모적방법；기우영향망락적태세고계모형，재계승전통패협사망락결구적기출상，이용촉진여억제량충인과영향관계사건모경위직관、역동，경호지표현증거절점대기자절점적영향관계；이용CAST라집사조건개솔표적획취경위간단，해결료조건개솔수절점증장이지수증장적문제，병능흔호지표현전가지식이급위조건개솔표적획취시제공료일충신적해결방법。
In order to overcome the difficulty caused by using Bayesian network in situation awareness in subjective situation and facilitate the excessive work to specify the CPTs the amount of which grow exponent with the number of nodes in the Bayesian network,a new method was proposed based on the influence net for situation awareness. The method proposes here uses two kinds of influence relations included promote and inhibiting influence,providing an intuitionistic and intelligible model for the situation awareness. By using CAST logic,the CPTs easily is gotten,the knowledge of domain experts is expressed,and a new method to is provided get CPTs when its difficult to get.