条件事件代数在数据融合中的应用

近年来,数据融合的研究集中到包括传感器管理、态势评估和威胁估计的较高层次。相对于较低层次的数据融合而言,这些研究领域更多地涉及到事件条件性和基于规则的证据的处理。条件事件代数是在确保规则概率与条件概率相容的前提下,把布尔代数上的逻辑运算推广到条件事件（规则）集合中得到的代数系统。文中介绍了条件事件代数的基本思想和一些性质,给出了利用条件事件代数实现人工智能专家系统中规则局部触发的例子。

Conditional event algebra and its application in data fusion

In recent years, much of the research emphasis in data fusion has shifted to the "higher" levels including Sensor Management, Situation Assessment and Threat Assessment which rely even more heavily on contingent or rule-based evidence than the lower level. Conditional Event Algebra (CEA) seeks to bring conditionality and contingency under the umbrella of standard probability theory. The basic ideas of Conditional Event Algebra are introduced, in which conditional events are defined as events in the infinite product of the original probability space. As an example, partial firing of rules in experts system is realized by using modus ponens modified by CEA.