一种知识一致性分析的非逻辑方法

A Kind of Nonlogical Method for Knowledge Consistency Analysis

知识的一致性分析是知识系统中的一个重要问题,是知识系统正常运行的基础。常规的知识一致性分析采用逻辑分析的方法进行,但不能完全解决实际系统中的知识不一致的问题,主要原因是严格的逻辑系统对满足约束条件的知识无法进一步分析。本文中提出的基于语义贴近度和多元线性回归模型的知识一致性分析方法正好弥补了这一缺点。在文章中给出了语义贴近度的定义及相应的算法设计。

The knowledge consistency analysis is one of the most important steps in the knowledge based systems,and it is the foundation for system to run.As a rule,we take the knowledge consistency analysis in a logical method.But the logical approach can not solve all the problems.In this paper,we present the knowledge consistency analysis method and algorithm based on the semantic proximity of attribute value and Multivariate Linear Regress model as complement to sovle the problem.