集值信息系统中基于证据理论的规则提取

Rule extraction in set-valued decision information system based on evidence theory

为了提高在集值决策信息系统中最优广义决策规则的获取速度,借助证据理论这一处理不确定性问题的有力工具,将证据理论应用到集值信息系统中,定义了集值信息系统在相似关系和优势关系下的基本概率分配函数,提出一种基于基本概率分配函数的最优广义决策规则获取方法。最后用实例说明了这种方法的有效性,并且有效地降低了时间复杂度。

In order to improve access speed of the optimal generalized decision rules from set-valued information system, this paper use evidence theory as an effective tool in dealing with uncertainty questions. This paper discusses the applica- tion of evidence theory in set-valued information systems, defines basic probability assignment functions on similarity rela- tion and dominance relation in set-valued information systems, proposes a new approach to extract optimal generalized deci- sion rules based on basic probability assignment. An example is given to prove the validity of this approach and it also shows the less time complexity of this approach than other methods.