Design And Implementation Of Intelligent Document Information Retrieval System For Government Affairs

With the development of government informationization,the new generation of information technology represented by big data,cloud computing and artificial intelligence has been widely used in the field of government field.At present,although the development of science and technology has greatly promoted the construction of government informationization,in the field of government document retrieval,how to effectively organize and use the document information with wide sources,plentiful types and complex structure is still a challenging problem.Therefore,this paper proposes to introduce the knowledge graph technology into the field of government field,and intuitively display the relationship between the retrieval results in the form of graph,so as to improve the user's search experience and the retrieval efficiency of official documents,and help reduce the workload of government information disclosure.The main work of this paper focuses on the following three aspectsFirstly,according to the requirements of the current government informationization construction of Guizhou Provincial General Office,the structure of the knowledge graph of official documents is defined.The required entity and attribute information is parsed from the document data source by rule-based method,and the knowledge graph is obtained after knowledge fusion.Considering the performance,neo4 j is used to store the knowledge graph.Then,for the complex,irregular document data,the method based on deep learning is used for entity extraction.In this paper,three models,Bi LSTM-CRF,BERT-CRF and BERT-Bi LSTM-CRF,are selected for comparative experiments.The experimental results show that the effect of entity recognition is significantly improved after the introduction of Bert pre training model.Therefore,the BERT-Bi LSTM-CRF is selected as the algorithm model for entity extraction of complex official documents,which provides a theoretical basis for the construction of high-quality official documents knowledge graph.Finally,based on the official document knowledge graph,using B/S development architecture and the flash development framework to achieve the document intelligent information retrieval system,completed the requirements analysis,system design,development and testing process.After testing,the developed system has perfect functions and stable operation,which can effectively improve the experience and efficiency of information retrieval.