| Agricultural scholars are important driving forces in the research,production,dissemination,and application of agricultural technology.With the development of the Internet,an increasing number of farmers and agricultural researchers are using search engines to obtain information on agricultural scholars and research results to meet their agricultural technology information needs in learning and production activities.This means that organizing and utilizing relevant information and literature results of agricultural scholars through digital technology not only facilitates the dissemination and sharing of agricultural technology information,but also has a positive significance for improving the efficiency and quality of agricultural production.This paper takes agricultural scholars in the Yangtze River Delta region as an example to conduct research on the construction and application of a scholar knowledge graph,which mainly includes ontology construction,information extraction,and topic discovery.The specific research work is as follows:(1)Ontology modeling of agricultural scholars.Firstly,based on the concept system of agricultural scholars and the key attributes of academic literature,the design of the ontology framework was completed.Secondly,the Protégé tool was used to define the class and property relationships,and the ontology was visualized,establishing an ontology in the field of agricultural scholars.Finally,based on the OWL language,the ontology was described and stored,and it was associated with existing information organization resources to lay the foundation for subsequent research on scholar information extraction.(2)Research on information extraction methods for agricultural scholars.In this paper,the BERT pre-training language model was used to represent the text data of scholars as word vectors,which were then input into the Bi LSTM-CRF model for named entity recognition experiments on agricultural scholars’ information.The performance of the domain pre-trained BERT model in entity recognition was compared with different vector input models and machine learning methods,and it was proven to be the best.Additionally,regular expressions were used to automatically extract metadata information from scholar literature,and the basic construction of the agricultural scholar knowledge graph was completed.(3)Evolutionary analysis of research topics of agricultural scholars.Firstly,the advantages and disadvantages of two document representation methods,CBOW and TF-IDF,were compared,and the TF-IDF algorithm was chosen for document representation.Secondly,the LDA topic model was used to extract the topic distribution of scholar literature,and the topic strength algorithm was used to generate the evolutionary trend of scholar research topics.Finally,we conducted theme extraction and evolution experiments with literature data of agricultural scholars in the Yangtze River Delta region,and analyzed the theme extraction results and evolution trends to expand the knowledge graph of agricultural scholars.(4)Agricultural scholar information service system was built based on Spring Boot,Vue framework,and Neo4 j database.The system deployed information extraction and topic discovery algorithms,and completed functions such as knowledge graph construction,scholar information query,and scholar portrait.A prototype of the agricultural scholar information service system based on knowledge graph was implemented. |
