Research On Domain Expert Identification Based On Knowledge Supernetwork

Driven by big data,data is growing at an exponential rate,showing a trend of becoming bigger,more complex,and more difficult.Academic resources are growing at a rate of double every two years,resulting in a complex knowledge network.Faced with massive academic resources,people inevitably fall into the dilemma of information overload and knowledge trek.Therefore,people pay more and more attention to the study of knowledge networks.People hope to explore the relationship between knowledge elements and the relationship between scientific entities,discover the knowledge associations in the knowledge network,better describe the knowledge network,and reveal the connections among entities in the knowledge network,so that we can achieve deep mining,integration analysis,knowledge discovery,and visual display of knowledge resources.Knowledge networks have the characteristics of hierarchy,complexity,and multiplicity.Therefore,general networks cannot describe complex knowledge networks.As a kind of heterogeneous network,super network is composed of heterogeneous nodes and heterogeneous relationships,and has properties of multi-level,multi-dimensional,multi-relational and nesting.Super network can more comprehensively describe the characteristics of knowledge networks,reveal the complex relationships of knowledge networks,and dig out the implicit relationships in knowledge networks,which is more consistent with the real knowledge network.As the main body of knowledge in the knowledge network,the experts who master the domain professional knowledge are important nodes in the knowledge network and even more valuable academic resources.In the increasingly competitive environment,we enter the era of talent competition,having a strong talent team,is the key element to quickly grasp the core technology to achieve innovation and progress.At present,expert resources are scarce,the national government,enterprises,universities and scientific research institutions find it difficult to select experts,and the evaluation mechanism of experts is not scientific enough.Based on these problems,how to objectively identify the domain experts with professional knowledge and meet the needs is a problem worthy of further study.In today's knowledge society,effectivelyidentifying experts in the field is an embodiment of precise knowledge demand,as well as a hot research in the field of information retrieval,knowledge management and service.Through literature review,this paper summarizes the current research status of super networks,and emphatically introduces the theoretical basis of knowledge super networks from the aspects of concepts,components,models,characteristics and applications.At the same time,this paper expounds the research status of the subject identification by domain experts;On the basis of the previous knowledge super network model,this paper expands the hierarchy of knowledge super network,constructs sub-networks from four dimensions of author,literature,domain and topic,describes each sub-network,and then elaborates the mapping relationship among each layer of networks,thus building a four-layer knowledge super network model;This paper uses the measurement index of super network,measures the literature influence degree in the literature sub-network,uses LDA model in the domain sub-network to extract domain tags and calculate the similarity among domains,uses TF-IDF algorithm in the subject sub-network to calculate the weight of the topic tag,and uses the vector space model to calculate the similarity degree of super-edge.Based on the idea of PageRank sorting,the super-edge sorting algorithm is proposed,and the domain expert recognition is carried out.And we propose the domain expert recognition mechanism based on knowledge super-network;Finally,the author takes the field of library and information science as an example,selects some papers of core journals in this field,constructs the knowledge super network model,and implements the algorithm mentioned above through experiments to identify the domain experts based on the experimental data.At the same time,the author compares the experimental results with h-index,p-index and social network analysis method,verifies the rationality of the domain expert identification algorithm based on the knowledge super network proposed in this paper,which provides new ideas and methods for expert identification and scholar evaluation.