| In recent years,with the rapid progress of the government information publicity,more and more governmental information resources are published in the governmental website.The information involves many fields,concerning national economy and the people's livelihood,so the masses have extensive demand.However,due to the hierarchy of the governmental organization,and different subjection and different division of the administrative body,the public information issued by the governmental website is characterized in hierarchy,regionalism and dispersity.People are always helpless faced with the abundant,insolent and disperse fragmental information,which obstruct the masses to get effective information and affect the governmental public information to be effectively used.The rapid growth of the information resources in the governmental website bring opportunities and more challenge meanwhile.Therefore,faced with the standoff between“information ocean” and “knowledge deficiency”,it is urgent to adopt certain methods and means to collect,recombine and integrate the multi-style information from multiple sources in the governmental website to build an orderly and highly-effective governmental information knowledge system and provide users to effectively get and use the knowledge.Faced with such problems like the fragmentization,redundancy and imperfection of the information,knowledge fusion can extract relevant knowledge from the multiple data sources to produce new integrated knowledge object through knowledge transition,integration and combination,which can build effective knowledge resources to solve the problem of decision-making.It provides new idea,new methods and new ways to solve the knowledge service of the information resource field of the governmental website.Starting from the realization of deep exploration and utilization of the governmental website information resources,with the goal of high efficiency of knowledge service for the masses,this essay comprehensively applies multiple theoretical methods including information management,computer sciences,knowledge sciences and system sciences tobuild the knowledge fusion framework model from the angle of knowledge organization and knowledge service,trying to change the fixed procedure of knowledge acquisition and form the new model of the process of knowledge acquisition.It proposes the theoretical framework and solution for the knowledge fusion for the dispersed and submersed governmental website information resources from multiple sources.On this basis,with the support of the semantic information organization technologies like semantic net,knowledge ontology and knowledge discovery,it tries from the three different dimensions including content dimension,structure dimension and application dimension to explore the realization method of the knowledge fusion for the governmental website information resources.The specific research contents include:(1)Through the explanation of relevant concepts like the governmental website information resources and knowledge fusion,it further clarifies the scope of research object and research content.Meanwhile,it theoretically traces to the sources of the knowledge space theory,knowledge organization theory and ontology theory and knowledge discovery theory,all of which have important guiding significance and reference value for this research.The multi-dimensional knowledge classification and knowledge management in the knowledge space provide important theoretical support for the construction of multi-dimensional knowledge fusion framework of the governmental website information resources.Knowledge organization theory has important guiding function in realizing the processes of knowledge representation,knowledge recombination,knowledge correlation and knowledge agglomeration for the governmental website information resources.Ontology is a great tool for the knowledge fusion of the governmental website information resources,while the knowledge discovery theory provides the technical method for the knowledge fusion of the governmental website information resources.(2)It confirms the goals and principles of the knowledge fusion for governmental website information resources.It analyzes the requirement of the knowledge fusion from the two aspects of the information requirement from users and the function requirement for knowledge fusion.It makes clear of the procedure for knowledge fusion and the basis for the construction of system structures.Then it analyzes the relationships between theconstitutional elements,functional elements and elements involved in the knowledge fusion process for the governmental website information resources under the semantic environment and proposes the logical framework of the knowledge fusion for the governmental website information resources,answering the question of “what to do” for the knowledge fusion for the governmental website information resources from the concept and logic levels.On the premise of the logical framework of the knowledge fusion for the governmental website information resources based on semantics,depending on the multi-dimensional knowledge category and knowledge management in the knowledge space,the classification of knowledge types in the knowledge fusion research field and the structure and functional framework of the multi-source information fusion system in the information fusion research field,it extracts the knowledge fusion function structure constituted by data-level fusion,concept-level fusion and decision-making level fusion.And it respectively builds the multi-dimensional knowledge fusion framework for the governmental website information resources from the three dimensions including content dimension,structure dimension and application dimension,answering the question of “how to do” for the knowledge fusion for the governmental website information resources from the structure and function levels.(3)The core problem in need of solution for the data-level fusion in the content dimension is how to make effective data correlation for the unified governmental website information resources to form knowledge network.For this problem,through fine information extraction and pre-disposal for the governmental website information resources in specific theme,it makes unified semantic description by depending on the meta data model motivated by ontology to build the knowledge ontology for the domain of governmental website information resources.By reference of the semantic network representation method,it realizes the knowledge representation for the governmental website information resources.On this basis,it introduces concepts of knowledge element and knowledge element-ontology view.Relying on the data correlation method based on the knowledge element semantic connection,it realizes the unified objectification representation for the knowledge element,the extraction of the knowledge element object,the nature acquisition of the knowledge element object and semantic correlation calculation of theknowledge element.And it accordingly builds the knowledge network for the macro description for resources content in specific theme.(4)The key problem in need of solution for the concept-level fusion in the structure dimension is how to obtain domain concept from the knowledge network of the governmental website information resources produced from the data-level fusion and identify the relations between domain ontology concepts to form the classified knowledge organization system convenient to discover the correlation rules and modes.For this problem,the essay proposes to apply the ontology formalization combined with RDF link qualitative reasoning and Voronoi quantitative reasoning to obtain the domain concepts of the governmental website information resources.Through identifying the domain ontology concept relations,it can build the hierarchy structure of the domain ontology concept relations to obtain the classified system for domain ontology concepts.On this basis,design domain ontology mapping model based on concept classification.Depending on the similar calculation method between ontology concepts,it can realize the knowledge ontology integration for the governmental website information resources based on the concept classification to eliminate the redundant knowledge in the knowledge network of the governmental website information resources to realize the goal of effective classification and organization.(5)The key problem in need of solution for the decision-making level fusion in the application dimension is how to make correlative classification for the correlative and similar domain knowledge in the knowledge network of the governmental website information resources and the classified knowledge system structure to form complete domain knowledge for decision-making application.For this problem,through comparative analysis on the existing method of correlative classification for domain knowledge,this essay proposes the fusion method for domain knowledge based on correlative extraction and semantic classification.This method involves 3 core fusion calculation methods for domain knowledge including timing sequence correlation fusion,cause and effect correlation fusion and agglomeration hierarchy fusion to form complete domain knowledge to assist users in decision-making.(6)By successively connecting the structural and functional classification of the multi-dimensional knowledge fusion framework of the governmental website information resources,the above part proposes the methods required in realizing the knowledge fusion in different dimensions.The effect of these methods still needs to be verified through experiments.Therefore,this essay establishes the software and hardware environment needed for the experiments and introduces relevant tools needed for the experiments.It describes the distribution situations of the experimental data sources and theme and clarifies the specific experimental contents for the knowledge fusion in different dimensions.After corresponding experimental verification,taking the domain information resources relevant to the theme of “Environmental Pollution and Protection” issued by the governmental website,it establishes the knowledge ontology example of “Haze Prevention” in the domain of“Environment Pollution and Protection”.Depended on the domain knowledge ontology of“Haze Prevention”,it respectively makes experimental verification for the knowledge network realization method based on the knowledge element semantic linkage proposed by the data-level fusion,the domain knowledge ontology combination method based on the concept classification proposed by the concept-level fusion and the domain knowledge fusion method based on correlation extraction and semantic classification proposed by the decision-making level fusion.Then it analyzes and evaluates the experimental results through designing qualitative and quantitative evaluation indicators. |
PDF Full Text Request