The Research On Organization Model Of Scientific Data Based On Linked Data

Scientific experimental data is an important component of the scientific data.Compared to the data in big science which have established well descriptive standards,data come from small science is facing some serious problems,such as lacking of preservation requirements and related policy support.In this circumstance,it is very difficult to re-analyze or to re-use those data to validate some experiments in current research.So there is a significant need for building a scientific experimental data organization model to keep the experimental data reusable and renewable.Firstly,the article constructs metadata describing model of botany gene expression by user interview.And the Delphi method was introduced to revise this model.Then the article explores the semantic relationship between metadata models and metadata sets through ontology technology.And finally,it realizes the data transforming to RDF format and linked data publishing by Linked data technology which supports to access linked data,seamless link for Web information and the retrieval of linked data based on SPARQL.The lightweight and multidimensional semantic network of science experimental data is successfully built in this article.The article achieves the goal of organizing the Botany gene expression experimental data by following three steps:(1)Building the experimental metadata model of botany gene expressionFirstly,the paper summarizes the current status of scientific experimental data organization by literature research,and then investigates the status of organization,the requirement of organization on gene expression experimental data through expert interviews to build the description model.Finally,the description model was established by using Delphi method.(2)Building the ontology of botany gene expression experimentIn the article,it discovers,mines and builds the semantic relationship between classes,entities and attributes based on semantic web.Besides,the article integrates and reuses the current ontology such as FOAF,DCMI,BIBO,EXPO and so on.And also,through the ontology building technology and the ontology building software Protege4.3.0,it builds experimental ontology of botany gene expression.(3)Building and publishing the linked data of experimental data.Writing the mapping file through D2R so as to bind scientific data with URI,the article publishes the linked data sets successfully.This makes it possible to organize,link and access different scientific data,so well as the retrieval and directional browsing based on SPARQL.Meanwhile,it supports the link and access between local scientific data and outer data such as scientific achievements,authors,relevant institutions and so on.The main innovation of the article is comprehensively applying multiple technologies to design and complete the scientific experimental description model on the basis of user interviews and Delphi method.The article constructs primary model through user interviews to know organizational status and demand of scientific experimental data.Then,Delphi method is used to score,select and revises the metadata model.This metadata model can meet the real demand of users in different domains,and also can follow the scientific data organization habits of scientific researchers.Knowledge supporting from experts ensures the description metadata model scientifically and professionally.