Research On The Semantic Aggregation Of Linked Named Entity Data

Linked named entity data use the data publication form of linked data,which indicates the alienation data of various entities,including the subject and external characteristic information of the corresponding entities.It is great value of reshaping the knowledge system,which contains rich semantic and complex correlations.At present,many semantic knowledge bases are based on open and linked for the construction of entity data such as individuals,organizations and places.In the same knowledge base,there is a strong correlation between entities represented by different named entity data,and their potential relationship is often not only for the user to display a single association,resulting in the loss of multidimensional features of the data and the reduction of data value,seriously affecting the overall level of knowledge base and data quality.In different knowledge base,the same entity object in construction methods,forms of expression,description of the scope and reveal the depth of the obvious differences,resulting in heterogeneous,sharing the degree of difficulty,low data utilization,which revealing the information ecological imbalance problem,such as information overload and information pollution,and which aggravate the cognitive burden of users in the network environment.Semantic aggregation has become an effective way to solve the above problems.It not only dynamically correlates and organizes knowledge fragments,provides clear directions and ideas for discovering new knowledge,but also eliminates the differences between with heterogeneous data,which can reorganize the organic and compact data aggregation model to meet the diverse needs of knowledge and services.In this article,we study the data semantic aggregation problem with theoretical analysis and empirical research,which based on homologous data and cross-source data.The research mainly focuses on the following aspects:(1)Analyzing the overall construction of named entity data in several typical knowledge bases,which respective compared their advantages and characteristics,providing realistic requirements for clarifying the problem of semantic aggregation.On this basis,the connotation and characteristics of the named entity data of the related entities are clarified,and a general association model of the names of the related entities is proposed.(2)Summarizing the realization method and application field of semantic aggregation,and we use it as the theoretical basis of semantic aggregation research,discuss the two polymerization bases selected in this paper,using the data linked to realize the homologous data aggregation,utilizing the semantic of data to implement cross-data source aggregation,the overall framework of semantic aggregation is designed accordingly.(3)Based on the chain solving method and the association rule minning technique,the homology semantic polymerization experiments using single person relationship and multiple entity relationships are realized respectively based on the datasets of the four major family figures in the Republic of China and the Nobel Prize laureates.For the improvement of GADES similarity measure method and Levenshtein Distance algorithm,two Nobel Prize winner datasets from Wikidata and YAGO are adopted as the object,which realize the semantic cross-source matching and aggregation implementation.Provide a refernce for the characteristics of cluster data mining,establishing linking,discovering resources and eliminating heterogeneous data sources.