论文部分内容阅读
为了缩减知识推理空间,提高分布式环境下知识处理的效率,提出分布式概念格属性约简的理论框架.基于粗糙集理论的思想,从子形式背景和全局形式背景的角度,刻画了核心属性、相对必要属性和绝对不必要属性的属性特征,给出属性约简的判定定理.在此基础上,给出概念格的分布式属性约简方法:首先,使用现有的约简方法分别计算各子形式背景的约简,然后,逐一利用各子背景的约简,通过合并计算得到全局形式背景的约简.给出了算法的实现并用实例验证了它的有效性.分布式约简有效避免了使用现有方法而引起的数据安全和网络通信等问题,提高了约简的计算效率.
In order to reduce the space of knowledge reasoning and improve the efficiency of knowledge processing in distributed environment, a theoretical framework for attribute reduction of distributed concept lattice is proposed.Based on the theory of rough set theory, the core attributes are described from the perspective of subforming context and global formal context , The relative essential attributes and the attribute attributes of absolutely unnecessary attributes, the judgment theorems of attribute reduction are given.On the basis of this, the distributive attribute reduction method of concept lattice is given.Firstly, the existing reduction methods are used to calculate The reduction of each sub-form background is then used to reduce the global-form background by using the reduction of each sub-background one by one. The implementation of the algorithm is given and its validity is verified by an example. Avoiding the problems of data security and network communication caused by using the existing methods, and improving the calculation efficiency of reduction.