因素空间中的背景分布与模糊背景关系

Background distribution and fuzzy background relation in factor spaces

背景关系在因素空间的因果分析理论中是一个重要的概念,由它决定了因素之间的因果联系.但是在现实生活中,背景关系往往带有不确定性,由之而产生的推理也具有不确定性.为了适应人工智能对处理不确定性信息的根本需要,本文将针对随机性和模糊性的特点而提出因素空间的背景分布和模糊背景关系.背景分布是因素相空间中的概率分布,基于背景分布,可以得到各个因素的边缘相分布和因素之间的条件相分布,对于事物因果分析提供了一种概率推理的方法.模糊背景关系的引入更对事物的因果分析提供了一种模糊推理的方法.

Background relation is an important concept in causality analysis of factor space theory. While in practical applications, the background relation is not determinate, but non-determinate, and it causes the uncertainty of inferences, In order to adapt to the fundamental needs to deal with uncertainty information in artificial intelligence, the paper presents background distribution and fuzzy background relation in factor space with respect to randomness and fuzziness respectively. Background distribution is the probabilistic distribution in phase space, based on which, we can get the marginal and conditional distributions of factors in phase spaces and get a way of probabilistic inference on causality analysis~ Similarly, fuzzy background relation provides a way of fuzzy inference on causality analysis.