Research On Aggregation And Service Recommendation Of Digital Library Resources

Under the circumstance of big data,digital library,which is characterized by digital storage and presentation,network retrieval and acquisition,becomes a kind of knowledge distribution center with the functions of knowledge acquisition,knowledge dissemination and communication.All kinds of digital objects(knowledge as an object,can be represented as a kind of documents,fact/research data,database,knowledge base and knowledge system)constitute a fully integrated digital library in the field of large data environment.The scale growth of digital library resources is irreversible.With the rapid increase of digital library resources and the popularity of intelligent technology,the intelligent and personalized recommendation service has become the trend of digital resource service recommendation.However,the traditional personalized service has been unable to meet the needs of users and the development of digital library.On the one hand,users' demands become more and more complex,so that how to accurately describe and fully tap the user's personalized complex needs and recommending resource service according to users' requirement become a problem;on the other hand,the number of digital resources is huge,so that how to effectively aggregate the massive resources,support efficient retrieval and recommendation,and fully tap the internal semantic relations of digital library resources become the focus of attention.The "aggregation" of digital library resources aims at discovering the semantic relationship between resources and their internal through the system convergence,mining and utilization,effective integration of multi-source heterogeneous digital library resources,constructing a resource system with interrelated,multi-dimensional and multi-level contents and forming a whole three-dimensional knowledge network made up of the concept of the theme,subject content and research object;As an effective intelligent information filtering technology,the service recommendation can combine the content characteristics of digital library resources and through the analysis and mining of the users' interest preferences and resource access behavior,it is able to recommend the digital library resources to meet the needs of the users.It can be seen that the service recommendation based on the aggregation of digital library resources has become an important way to meet the needs of the users.Based on the techniques and methods such as domain ontology,data mining,social network analysis and so on,learning from the results of polymerization and personalized recommendation,this dissertation presents a method of resource aggregation and service recommendation for users' complex requirements,providing reference for digital resource service recommendation.The main research contents are as follows:(1)Define the concept of digital resource aggregation and service recommendation and expound the relationship between the two parties.The aggregation of digital library resources is the basis of the service recommendation and the service recommendation is the goal of the digital resource aggregation.Data aggregation,information aggregation and knowledge aggregation quality are the key to the service recommendation.Service recommendation guides the quality of aggregation as well from users requirements,which further reveals the knowledge structure and development law of subject field.Thus,it can meet the user's individual needs and improve the service recommendation effect on the basis of improving the aggregation ability and efficiency of digital library resources.(2)Describe the concept of the digital resource semantic,analyze the semantic distance and semantic of digital library resources,and the concept of digital library resources,starting from the perspective of Digital Resource Semantic and semantic relations.Semantic concepts and semantic relations constitute the basis of the aggregation of digital library resources,and regard the domain ontology as a directed network composed of nodes(semantic concepts)and arc chains(semantic relations).Analyze it based on complex network methods and deeply reveal the network relationship between concepts,furthermore,dig,establish and make full use of the connection between the digital library resources and show the relevance through aggregating the complex and disorderly content.(3)Systematically study the characteristics and dimensions of the aggregation of digital library resources,use the tools of social network analysis and semantic similarity calculation and construct the aggregation model of digital library resources.The user is the subject of digital resource aggregation and service recommendation,and domain ontology is the most effective theory and technology of the digital library resources from the perspective of knowledge organization structure.Based on this,the characteristics of the aggregation of digital library resources are summarized:diversification of digital resource aggregation objects,diversification of digital resource aggregation methods,dimension of the aggregation of digital library resources and dynamic process of the aggregation of digital library resources.On this basis,the digital resource aggregation model is proposed.(4)Analyze the network structure of the domain ontology based on Wikipedia using social network analysis and put forward an important node identification method of domain ontology network combined with the measure of the degree of social network analysis and clustering coefficient,as the basis of polymerization.Based on the D-S evidence theory,the recognition framework(high,low)to construct the basic probability assignment function,said node importance,are merged by D-S theory of evidence combination formula,in the form of probability to quantify the importance of nodes,and the node importance evaluation index.Using the social network analysis method to construct the domain ontology network and exploring important nodes in the network can provide the method and reference for fully exploiting the semantic information of ontology,finding the implicit knowledge in the ontology,analyzing and presenting the knowledge structure of the specific domain,and sharing the domain knowledge.(5)An improved spectral clustering algorithm is proposed by improving the semantic similarity computation method based on ontology and the semantic similarity calculation method based on the synonym expansion.Use a method of text clustering analysis in the aggregation of digital library resources and start from the internal and external characteristics of digital library resources and use the principle of resource aggregation to make the digital library resources clustered.Then,one or more suitable representation structures are selected according to the hierarchical cluster structure of the digital library resources,so as to dig out the semantic relationship between the digital library resources.(6)Propose 3 service recommendation methods:digital resource content service recommendation method based on ontology rule reasoning and semantic similarity calculation,service recommendation method based on Association semantic chain and service recommendation method based on spectral clustering.Through the effective aggregation of the digital library resources,the resources are semantically and semantically connected.On the basis of this,the user's demand information is excavated and the user's preferences are recommended.By Using the methods of ontology,data mining and service recommendation and researching on the aggregation of digital library resources and service recommendation,a new framework for resource aggregation and service recommendation is proposed,so as to provide valuable reference and guidance for optimizing the resource re organization structure and improving the knowledge service ability.(7)Using HowNet literature resources as a data source,verify the polymerization and recommendation method proposed in the fourth chapter and the fifth chapter.From the view of technique,JAVA language and MYSQL database are used to design and implement the semantic retrieval and application platform of digital library resources.The system uses the C/S system architecture,which integrates the aggregation method and the recommendation method proposed in this dissertation and achieves the function of digital resource aggregation and service recommendation.The method of Service recommendation based on aggregation solves the problems of traditional recommendation methods,such as insufficient mining user requirements,cold start and data sparseness,achieve a high degree of recommendation results and the user's personalized needs and improve the recommendation accuracy,then further expand the new vision of the future development of knowledge service in digital library.