摘要
针对知识库的建立需要耗费大量的时间和人力,同时相同或相似领域的知识库数量越来越多,提出利用现有规则知识库进行合并生成一个新的规则知识库,并对生成的新规则知识库进行知识冗余、环路和冲突的检测算法。首先,规则库利用有向超图来表示;其次,将有向超图利用其邻接矩阵来表示,那么规则库的合并可以转换成有向超图所对应的邻接矩阵的合并,并依据邻接矩阵求可达矩阵以及利用总可达矩阵来检测规则库中规则的冗余、环路和冲突。最后,算法的有效性通过实例加以验证。
To- build the knowledge base needs to spend a lot of time and manpower, meanwhile there are more and more knowledge bases built in the same or similar fields. In light of this, we proposed an algorithm, which utilises existing rule knowledge bases and merges them to generate a new rule knowledge base, and detects the knowledge redundancy, circularity and conflict in the generated new rule knowledge base. First, we represented the rule base by the directed hypergraph; secondly, we represented the directed hypergraphs by their adjacency matrices, thus the merging of rule bases can be converted into the merging of the corresponding adjacency matrices of the directed hypergraphs, we then obtained the reachability matrix according to adjacency matrices, and used the total reachability matrix to detect the redundancy, circularity and conflict of rules in rule base. Finally, the effectiveness of two algorithms was verified through examples.
出处
《计算机应用与软件》
CSCD
2016年第1期249-253,311,共6页
Computer Applications and Software
基金
国家自然科学基金项目(31171456
61203373)
中国科学院战略性先导专项(XDA08040110)
关键词
规则库
规则库合并
有向超图
邻接矩阵
可达矩阵
Rule base Merging of rule base Directed hypergraph Adjacency matrix Reachability matrix
