Retruieval And Visualization Of Scientific And Technological Resources For Experts And Scholars Based On Knowledge Graph

Institutions of higher learning,research institutes and other scientific research units have abundant scientific and technological resources of experts and scholars,and these talents with great scientific and technological innovation ability are an important force to promote industrial upgrading.The scientific and technological resources of experts and scholars are mainly composed of basic attributes and scientific research achievements.The basic attributes include information such as research interests,institutions,and educational work experience.However,due to information asymmetry and other reasons,the scientific and technological resources of experts and scholars cannot be connected with the society in a timely manner,and social needs cannot be accurately matched with experts and scholars.Therefore,it is very necessary to provide relevant expert and scholar retrieval services.The work of this thesis is mainly based on the retrieval and visualization of scientific and technological resources of experts and scholars based on the knowledge map,and the following tasks have been completed:(1)A Knowledge graph construction of scientific and technological resources for experts and scholars.Relevant technologies such as crawlers are used to obtain the scientific and technological resource data of relevant experts and scholars on the web page,and the joint extraction method of entities and relationships is used to realize the construction of the knowledge graph of the scientific and technological resource information of experts and scholars.In the method of obtaining scientific and technological resources data of experts and scholars,regular expressions and BeautifulSoup module are used to parse the current web page to complete the automatic collection of relevant information.For Chinese text,a combined text relation extraction algorithm based on word mixing and GRU(MBGAB)is proposed,which combines word mixing vector mapping method,gate control unit and self-attention mechanism to achieve knowledge triple extraction.On the basis of the two,the entity and relationship types of the scientific and technological resources of experts and scholars to be extracted are predefined,and the MBGAB algorithm is used to complete the construction of the knowledge map of scientific and technological resources of experts and scholars.(2)Semantic representation learning of scientific and technological resources for experts and scholars based on knowledge graph.Aiming at the problem that pre-trained language models have poor semantic representation of texts in professional fields,a knowledge representation learning algorithm based on adversarial generative networks and pretraining(KRGP)is proposed.The adversarial generative networks are used to generate high-quality negative triplet samples,based on entity descriptions.Construct the text sequence corresponding to the triplet,finetune the pre-trained language model,integrate the language model into the professional knowledge of the knowledge map,and enhance the semantic expression ability of the pre-trained language model.Based on the triples of scientific and technological resources and knowledge of experts and scholars extracted,use Wikipedia to obtain their entity text descriptions to achieve semantic expansion,use the adversarial generation network to generate positive and negative samples,and splice the triples into the BERT input form.The Next Sentence Prediction(NSP)task fine-tunes BERT to obtain a language model that integrates the knowledge of scientific and technological resources of experts and scholars,and realizes the semantic representation learning of scientific and technological resources of experts and scholars based on knowledge graphs.(3)Text semantic retrieval of scientific and technological resources for experts and scholars.Aiming at the application scenario of related word retrieval,a semantic retrieval algorithm of scientific and technological text based on similarity calculation is proposed.The scientific and technological text is input into the language model to obtain the representation vector.At the same time,considering the nonlinearity and singularity of the vector distribution,the text is processed twice by linear transformation.Coding,calculate the similarity between the output vector value and the target keyword,sort according to the size of the similarity value,and return the search result.Construct an online semantic retrieval framework for scientific and technological resources of experts and scholars based on Elasticsearch,use the semantic retrieval algorithm of scientific and technological texts based on similarity calculation,and import the text semantic vectors of scientific and technological resources of experts and scholars into Elasticsearch,and then use cosine similarity to calculate the search terms and all texts The semantic similarity is used as a scoring function to return the results to be searched,so as to realize the textual semantic retrieval of scientific and technological resources of experts and scholars.(4)The experts and scholars’ scientific and technological resources retrieval and visualization system.The system includes four modules:the construction module of knowledge map of scientific and technological resources of experts and scholars,the learning module of semantic representation of scientific and technological resources of experts and scholars,the module of semantic retrieval and recall of scientific and technological resources of experts and scholars,and the query and visualization module of scientific and technological resources of experts and scholars.The system can search keywords from the massive data of experts and scholars’ scientific and technological resources,display the query results of experts and scholars through the network interface,and statistically analyze the relevant information in the scientific and technological resources of experts and scholars.