The Methods Of Information Service For Big Earth Data

The concept of Big Earth Data is emerging as a new inter-disciplinary integration of Earth science,space science,information science,computer science and data science.It utilizes the systematical and holistically approach for the association relationship research of the spatial data and geo-information,to discover the knowledge of the interaction of the Earth spheres.This research takes the information service as the way,through which serving knowledge discovery of Big Earth Data.A life cycle model for Big Earth data is established including data cycle,information cycle and knowledge cycle.The research is conducted from the content service for data and information and the process service for knowledge discovery.The research analyzes the connotation of information service,the construction method of information service system,metadata association method and the information service credibility evaluation method.The three outputs of this research are as the following:(1)This research studies the theoretical connotation and the methodology of the information service for Big Earth Data.It has solved the issues of the information service model for the knowledge discovery of Big Earth Data and the cloud service architecture for information service system.It is proved that it can provide fundamental information service platform for Big Earth Data,as a data service system in the cloud has been designed and deployed.The data loop,information loop and knowledge loop and the procedures of the Big Earth Data life cycle model are developed.Through a SWOT(Strengths,Weakness,Opportunities,and Threats)analysis,a holistic framework including model,technology and standard of information service for Big Earth Data is proposed.The service models for professional users,decision-making users and public user are analyzed,together with the 11 capabilities of the information service in the two service types as content service and processing service.It demonstrates that the system architecture of information services in the cloud could embrace a hybrid way of ?service-oriented architecture‘ and ?system of systems architecture‘.The approach of data categorization on the database cluster is used to design the system's data service architecture,and the approach of micro-service architecture is used to design the system's analysis service architecture.A comparative system performance analysis is implemented.The results show that the data online system can basically meet the requirements,and the method of system construction has the advantage of cost performance.(2)This research proposes a core metadata collection for Big Earth Data information service and a metadata association method.It has solved the issues of compatibility and associability of the metadata model for Big Earth Data.The method is proved that it can be beneficial to content searching and indexing for the knowledge discovery,as it provides the keywords for bridging the association relationship in terms of scientific objectives,application fields and service directions.The method is in accordance with the concept of ?core metadata,plus content keyword linking,plus semantic mapping‘.It adopts a ?package,entity and element‘ three-level hierarchical layer,and has spatial package,temporal package,feature package,status package and association package.A total of 16 entities and 42 elements in these packages are composed as elements for the core metadata.The UML(Unified Modeling Language)diagram and the data dictionary are also defined.Knowledge association,application association and service association through keywords are strengthened by employing the science and application traceability matrix method.Semantic mapping for enhancing the interoperability among the proposed core metadata and other international metadata standards is implemented.The core metadata enabled data service system on the cloud is developed,and the method is applied to knowledge discovery use case of lake area change.The study demonstrates the proposed metadata network method is able to offer the ability of integrating,publishing and indexing heterogeneous resources,based on the knowledge,application and service context.It also proves this research has the potentials to enhance the capacity building for knowledge discovery.(3)The research proposes a credibility evaluation method of the information service for the knowledge discovery of Big Earth Data,for the first time.It has solved the issues of the credibility evaluation method from the perspective of a life cycle.The method is proved that it can be beneficial to the validation of the contents and processes for the knowledge discovery,as it has developed an integrated evaluation procedure combined with the concept of data curation.The evaluation indicator system is built,from the perspectives of data quality,model and information availability and satisfaction degree of knowledge.The fuzzy mathematics comprehensive evaluation method and the analytic hierarchy process are used as the integrated measures for the evaluation.The thesis takes a lake size monitor use case to test the evaluation method.The proposed evaluation method is used for a case study of monitoring the lake changes in China with multi-dimensional time-series Earth observation data.The evaluation indicator system is built from the aspects of data integrity,data accuracy,data consistency,model availability and model result availability.There are totally 16 primary evaluation indicators and 31 secondary evaluation indicators proposed.The evaluation results are calculated from an evaluation matrix together with the weight of the indicators and the grade of the evaluation,and then the evaluation results are verified.It shows that this method can provide support for evaluating the credibility of Big Earth Data.The study demonstrates the proposed evaluation method is able to offer supports for measuring the credibility of Big Earth Data.It also proves this method has the ability to serve verifying the values for knowledge discovery.