基于Ontology的跨组织知识整合方法研究

A Study of Cross-Organizational Knowledge Integration Methods Based on Ontology

从知识整合的实际需求和研究现状出发,给出了跨组织环境下基于Ontology技术的知识整合流程,并结合语义网技术实现了知识集的闭包、查询与标准化同构等算法,开发了跨组织知识整合的原型系统。通过一个钢铁企业跨组织知识整合案例展示了原型系统的功能,说明知识的整合可以达到"语义增值"的效果;通过两个实验给出了原型系统的性能。

The prevalent MA business practices call for attentions to the effective management and use of inter-organizational knowledge for organizational learning.Inter-organizational knowledge sharing is more complicated than the traditional intra-organizational knowledge sharing because the former has a higher degree of knowledge inconsistency,misrepresentation,and inconsistency of terminology.The purpose of this paper is to improve the effectiveness of knowledge sharing across organizations by addressing those inconsistency issues. We first standardize definitions for knowledge resided in different organizations,and then integrate knowledge across organizations based on the ontology method.We further use the Semantic Web technology to develop algorithms for closure,query and isomorphism.The proposed knowledge integration process include the following activities：（1） users input the query conditions,（2） knowledge center inquires each individual corporation and concludes with a set of knowledge（3） knowledge center re-inquires and combines the results,（4） knowledge center computes the closure according to the public ontology,（5） knowledge center computes the closure＇s standardized isomorphism and provides feedback on representation module,and（6） the representation module organizes the results and shows the readable query result to the user. A prototype system is developed according to closure,query and isomorphism algorithms.A business case of knowledge integration among steel companies is used to show the functions of the system.The case shows that effective knowledge integration processes can produce incremental knowledge.We also conduct two experiments to measure the performance of the prototype system.The time cost increases almost linearly with the increasing scale of the knowledge base.In contrast,the time cost for the results remains stable with the increasing scale of target set.Effective knowledge sharing process can increase knowledge values. Knowledge integration is an important part of knowledge sharing.After standardizing ＂knowledge integration＂ definitions,we propose a complete set of mechanisms to integrate knowledge across organizations and use the Semantic Web technology to execute those mechanisms.This study proves the effectiveness and performance of our prototype system.