The Quality Analysis Of Interoperability Of Controlled Vocabularies

With the development of internet and big data,a lot of heterogeneous data emerged.As the basis of semantic extension,cross-database search and data sharing,interoperating among controlled vocabularies becomes an important way to realize one-stop information retrieval.Controlled vocabularies interoperation is a kind of semantic interoperation,which is accomplished through creating mappings between vocabularies using concepts and concept relations in order to make semantic links among vocabularies.In the process,we should choose appropriate methods and follow standard procedures according to the specific goals.At present,there are already established methods and models of interoperation.And interoperations have been established among heterogeneous vocabularies with the differences of languages,types and domains.In the aspect of interoperability standards,there are ISO 25964,ANSI/NISO Z39.19-2005 and BS 8723-2007 and other standards to guide the choosing of methods and models and solve the problems occurred in the process.However,the researches of problems and quality of interoperation are rarely.It is essential to analyze the quality of interoperation among controlled vocabularies,which contributes to discover problems occurred in the process,assess the effect of interoperate result,improve interoperate methods,reduce the loss of information,and eventually enhance retrieval efficiency.The paper analyzes interoperate quality systematically in terms of three aspects.First is the analysis of influence factors related with interoperate quality,summing up the influence factors from basic methods and vocabulary heterogeneous.Secondly,the quantitative analysis of interoperating results from an overall perspective,carry out a statistical analysis on the selected instance.The third is the analysis of retrieval effectiveness,analyzing the practical impact of vocabulary interoperation on retrieval results by a contrast experiment,which is between extended search based on UMILS Metathesaurus and basic search indexed by MeSH.The author draws some conclusions through the above analysis.The quality of interoperation result is different due to the different methods and models,mapping types,mapping directions,and mapping quantities.And vocabulary heterogeneous caused by structure,syntax and semantic is impact on the understanding of concepts and concept relations,which indirectly determines the select of methods in the interoperate process.And the quality analysis of interoperate result provides an overall view of mapping distribution.The result of retrieval experiment shows that the retrieval effectiveness of extended retrieval based on terminology mapping is not necessarily better than basic retrieval based on a single vocabulary.