关键词: 超网络模型, 语义搜索, 知识检索, 数据挖掘, 知识管理

Abstract: In knowledge-driven industrial software development,assisting engineers in searching heterogeneous semi-structured knowledge efficiently and accurately is a major issue.A semantic retrieval method was proposed based on the knowledge super network model.The knowledge super network consisting of product subnet,object subnet,and knowledge subnet was built with the relations between the concepts of code reuse and the attributes of engineering knowledge.To calculate the process context correlation between user query and engineering knowledge,the conceptual knowledge and language model were integrated by Bayesian method.Experimental results on Microsoft knowledge base dataset show that the proposed approach could improve the precision of knowledge retrieval comparing to several semantic retrieval methods.The feasibility and effectiveness of the approach were also verified.

Key words: hypernetwork model, semantic search, knowledge retrieval, data mining, knowledge management