基于“质”与“量”的船舶领域知识图谱构建研究

Research on Construction of Knowledge Graph in Ship Domain Based on Quality and Quantity

[目的/意义]为满足船舶领域科研或工程人员在知识问答、质量分析等方面的业务分析需求,提升科研工作效率与知识获取便利性、精准性。[方法/过程]围绕结构化、半结构化、非结构化等船舶领域多源异构数据,文章从质和量两方面提出了船舶领域知识图谱构建方法,基于船舶领域概念图谱与实体图谱构建,实现知识图谱“质”的描述;通过对实体间关联关系进行统计分析与建模分析,建立面向业务应用场景的量化分析模型;最终以问答推理为例,展示了从质和量两方面开展面向业务场景知识图谱构建的通用过程。[结果/结论]通过以舰艇涂料失效知识图谱构建为例,验证了在质和量两个方面构建知识图谱的正确性,该方法能够为业务场景分析提供有效支撑。

[Purpose/significance] In order to meet the business analysis needs of scientific research or engineering personnel in the marine field in terms of knowledge question and answer,quality analysis,etc.,to improve the efficiency of scientific research and the convenience and accuracy of knowledge acquisition.[Method/process] Focusing on structured,semi-structured,unstructured and other multi-source heterogeneous data in the ship field,this paper proposes a method for building a knowledge map in the ship field from both qualitative and quantitative aspects.Based on the concept map and entity map in the ship field,the knowledge map is realized.The description of “quality”;through statistical analysis and modeling analysis of the relationship between entities,a quantitative analysis model for business application scenarios is established;finally,question and answer reasoning are taken as an example to demonstrate the development of business scenario-oriented knowledge from both qualitative and quantitative aspects The general process of map construction.[Result/conclusion] Finally,taking the construction of the naval coating failure knowledge map as an example,the correctness of the construction of the knowledge map in both quality and quantity is verified.This method can provide effective support for business scenario analysis.