Construction Of Knowledge Graph For Government Service Convenience Hotline And Its Application In Intelligent Dispatching

With the rapid concentration of urban population,the imbalance of resource supply and demand has intensified,and the interests of citizens have been diversified.The government service convenience hotline,as a bridge for direct communication between citizens and the government,has become an effective channel for urban governance,and citizens’ trust in the hotline is increasing.Due to the large number of hotline access,especially in the event of public health emergencies such as the epidemic,the monthly order volume of more than 100,000 orders has become the key to restricting the effectiveness of services.The hotline intelligent dispatching system that supports users’ natural language input has become the key research object of urban intelligence in recent years.With the development of knowledge graph technology,it provides services for the intelligent dispatch of government services to facilitate the people’s hotline.In view of the construction of the government service convenience hotline knowledge graph and the application in the intelligent dispatch order,the work done in this article is as follows:(1)Build a government service convenience hotline itself.First,data cleaning,data analysis and other operations are carried out on the data of the government service convenience hotline,and then combined with ISO/IEC 5087-1 and ISO/IEC 5087-2’s two urban data model standards and government service convenience hotline data characteristics,and the TOVE enterprise modeling method is used to complete the government affairs service The construction of the ontology of the service convenience hotline,and finally the use of the open source ontology construction tool Protégé is finished to build the ontology of the government service convenience hotline.(2)It is proposed that the entity relationship between the fusion ALBERT(A Lite Bidirectional Encoder Representations from Transformers,ALBERT)and the pointer network annotation is jointly extracted.The corpus labeling changes according to the different needs of the data.This thesis uses the Albert pre-trained language model to complete the government service according to the data characteristics of the government service convenience hotline.The training of the service convenience hotline data,and the joint extraction of entity relationships is realized by using the pointer network annotation strategy,which not only improves the labeling efficiency,but also effectively alleviates the problem of overlapping relationship extraction.At the same time,the joint extraction model based on the entity relationship between the fused Albert and the pointer network is experimented on the data set,and 93.62% of the F1 value is obtained.First,the joint extraction method of the entity relationship between the fusion Albert and the pointer network is used,and the entity that is the hotline data according to the government service convenience hotline body,Relationships and attributes complete the extraction of the three-tuple data and import the triple into the Neo4 j graph database,complete the construction of the knowledge graph of the government service convenience hotline.(3)This thesis puts forward the intelligent distribution algorithm of government service convenience hotline based on knowledge graph.First of all,the entity recognition of different hotline scenarios is realized by using the Chinese word segmentation,part-of-speech annotation,dependency syntactic analysis and semantic character annotation functions in the LTP(Language Technology Platform,LTP)natural language processing tool to realize the entity recognition of different hotline scenarios.Then,the entity link model of BERT-CNN(Bidirectional Encoder Representations from Transformers-Convolutional Neural Network,BERT-CNN)is proposed to show the results of intelligent order dispatch in the scenarios of different government services and convenience hotlines.Finally,through the TF-IDF(Term Frequency–Inverse Document Frequency)similarity calculation to improve the accuracy of the intelligent dispatching system of the government service convenience hotline,realize the intelligent dispatch of the hotline.(4)Design and realize the government service convenience hotline intelligent dispatching system based on knowledge graph.This thesis combines the Django framework to realize the development of the dispatching system,realize the user input of government service convenience hotline data in the way of natural language,and realize the intelligent dispatch order of government service convenience hotline by matching common hotline scenario corpus.This work provides a new idea for the field of government service convenience hotline in smart cities,among which the knowledge graph of government service convenience hotline can provide a basic platform for the subsequent knowledge base of government service convenience hotline construction and intelligent inquiry of government service convenience hotline.