Research On Knowledge Service Model Of Enterprise Risk Management Based On Knowledge Graph

With the rapid changes of information technology,user demand and other factors,enterprises are facing increasingly complex and volatile environment.Due to the complex and changeable internal and external environment of enterprises,various risk factors are highly concentrated,which leads to frequent enterprise risk events.The relevant enterprises are restricted by the insufficient early warning and disposal ability of unexpected risks,and are faced with the risk of bankruptcy.With the development of digital economy globalization,enterprise data presents explosive growth,and big data environment provides massive data for risk management decision-making.However,these risk data present a fragmented and isolated organizational status,which makes enterprises unable to grasp the knowledge service demand in each stage of risk events from the perspective of systematization and intelligence,and unable toquickly realize the identification,early warning and processing of risks.The mature and application of knowledge graph technology provides an opportunity to solve the above problems.The application of knowledge element,ontology,knowledge fusion,knowledge mining and reasoning technology means enables us to construct multi-level and multi-dimensional knowledge organization form for risk big data,break the data island phenomenon in risk domain,form interrelated knowledge service network among risk elements,and carry out real-time monitoring and processing on the occurrence and evolution of enterprise risks.The knowledge service system of enterprise risk management based on knowledge graph can effectively organize and manage knowledge resources in the field of enterprise risk management,effectively improve the ability of enterprise risk identification,prediction and response,and assist managers to make accurate and efficient risk decisions.At present,there are few researches on knowledge service of enterprise risk management.Existing researches mainly start from the level of technology and method to innovate the knowledge service content of enterprise risk management and expand the data dimension in the field of risk management.These researches mostly focus on the risk management information service level,and lack of knowledge services such as risk identification,risk early warning and risk treatment embedded in the whole process of enterprise risk decision-making based on the overall needs of enterprise risk management.This will not be conducive to managers’ accurate acquisition of risk decision-making knowledge and seriously affect the efficiency of enterprise risk management and knowledge service quality.In view of this,based on the field of enterprise risk management,from the perspective of knowledge service,this dissertation uses the theory and method of knowledge graph,taking financial service enterprises as an example,to construct the knowledge graph in the field of enterprise risk management,solve the problems of knowledge representation,organization and processing in the field of enterprise risk management,and provide a complete knowledge system for the research of knowledge service in the field of enterprise risk management.At the same time,this dissertation will construct the enterprise risk management knowledge service model based on knowledge graph from the three aspects of enterprise risk identification,risk early warning and risk control scheme recommendation,and integrate artificial immunity and case-based reasoning and other related technologies to provide intelligent support scheme for enterprise risk management knowledge service system.The main research work and conclusions of this dissertation are as follows:(1)Design the theoretical framework of knowledge service for enterprise risk management based on knowledge graph.Under the guidance of complex system theory,this dissertation analyzes the elements and operation process of enterprise risk management system.Based on the knowledge demand of enterprise risk management,this dissertation analyzes the characteristics of enterprise knowledge service in the process of risk management,and puts forward the knowledge service mode oriented to risk management.By analyzing the elements of knowledge graph and the knowledge structure of risk domain from the perspective of knowledge service,the role of knowledge graph in enterprise risk management knowledge service is obtained,and the construction process of knowledge graph in enterprise risk management domain is given.On this basis,the elements and process of knowledge service for enterprise risk management are further explored,and a theoretical framework of knowledge service for enterprise risk management based on knowledge graph is designed to provide theoretical framework guidance for subsequent research on knowledge service.(2)Construct the knowledge graph for enterprise risk management domain.From the perspective of system,this dissertation aims to solve the practical problems of enterprise risk management,and takes knowledge service as the link.The entity extraction model with stroke ELMo embedded in IDCNN-CRF was used to obtain domain entities from unstructured risk domain text data,the entity relationship extraction model with mutual attention mechanism was used to accurately mine inter-entity relationships from cross-connected diversified entities,and the entity alignment model based on representation learning was used to integrate and recombine knowledge in knowledge extraction stage Group,use Neo4 j graph database to store complex graph structure knowledge in the field of enterprise risk management,use semantic path-based meta-learning model to complete and improve the knowledge graph in the field of enterprise risk management,methodically collate the knowledge organization process in the field of risk management,and build the knowledge graph in the field of financial service enterprise risk management.To provide a complete domain knowledge system for related enterprise risk management knowledge service research.(3)Construct enterprise risk intelligent identification model based on knowledge graph.Based on enterprise risk management domain knowledge graph,entity link is made between enterprise risk event description text information and risk entities in knowledge graph to obtain stronger risk characteristics.The enterprise risk feature system is constructed from the two dimensions of text vector and language feature of the core entity in the risk domain.Based on the theory and method of artificial immunity and the immune system as a reference,the entity characteristics of the risk domain are transformed into the risk antigen characteristics,and an intelligent enterprise risk recognition model based on artificial immunity is constructed.The model uses the improved V-detector algorithm to construct a risk discriminator for intelligent identification of domain text information,and carries out empirical analysis on the risk control data disclosed by enterprises and the enterprise network media data obtained by web crawler,realizing the efficient identification of enterprise risk based on knowledge graph.(4)Construct the enterprise risk automatic warning model based on knowledge graph.This dissertation analyzes the operation process of enterprise risk early warning system from four aspects: the selection of early warning index,the measurement of early warning index,the construction of early warning model and the prediction of risk information.According to the connotation of enterprise risk early warning and the principle of index selection,entity risk source and risk impact in the risk field were selected as early warning indicators,and the stroke ELMo model and semantic correlation degree were used to calculate the risk value of each indicator,and the risk early warning indicator system was established.On the basis of giving the domain entity risk value,the knowledge graph of enterprise risk management domain is empowered and expanded,the risk value of early warning index is endowed with the knowledge graph of risk management domain,and the risk early warning model based on the weighted knowledge graph is constructed.Taking financial service enterprises as an example,the comprehensive risk value of enterprise risk management domain text is calculated through the weighted knowledge graph.Realize the early warning analysis of enterprise risk.(5)Construct the enterprise risk control scheme recommendation model based on knowledge graph.This dissertation systematically analyzes the service objective,process and framework of enterprise risk control scheme recommendation,takes the knowledge graph of enterprise risk management domain as the knowledge basis,uses the triplet method to represent enterprise risk events,and combines the characteristics of enterprise risk events to put forward the enterprise risk management case retrieval strategy.The combination of CBR and RBR is used to build a recommendation model of enterprise risk control scheme,and empirical research is carried out on the model through specific enterprise risk events to verify the effectiveness of the model and achieve accurate recommendation of enterprise risk control scheme.Compared with domestic and foreign researches in this field,the research innovations in this dissertation are mainly reflected in the following three aspects:(1)Construct the theoretical framework of knowledge service based on domain knowledge graph.Based on the research results of knowledge service and the practical experience of enterprise risk management,and on the basis of analyzing the role of knowledge graph in knowledge service of enterprise risk management and its related elements,a three-layer architecture of "knowledge acquisition,knowledge organization and knowledge application" is proposed for enterprise risk management,which aims to meet the knowledge needs of enterprise risk management and constructs the theoretical framework of knowledge service based on domain knowledge graph.This theoretical framework breaks through the traditional knowledge service mode of "knowledge service resource-knowledge service process-knowledge service subject",deeply integrates the knowledge service objective based on knowledge graph with enterprise risk management needs,further optimizes and expands the approach and process of enterprise knowledge service,enriches the relevant theories of enterprise knowledge service,and provides a new model and theoretical system for the construction and application of enterprise risk management knowledge service system.(2)Construct the knowledge graph for enterprise risk management.From the perspective of the application of knowledge graph in enterprise risk management,this dissertation organically combines enterprise risk management with knowledge graph,innovatively puts forward the construction process and model of knowledge graph in enterprise risk management domain such as knowledge extraction,fusion,storage and reasoning,and builds the knowledge graph in enterprise risk management domain.It realizes standardized modeling and transformation of semantic knowledge in the risk management domain,effectively solves problems such as orderiness and relevance of knowledge in the enterprise risk management domain,and provides complete knowledge support for enterprise risk management knowledge service.This study provides a new perspective for enterprise risk management knowledge service system innovation and optimization of knowledge organization and knowledge service content and form.At the same time,based on the knowledge graph of enterprise risk management domain and combined with important scenarios such as enterprise risk identification,risk early warning and risk control plan recommendation,the application analysis is carried out to innovate and improve the enterprise risk management knowledge system,achieve efficient enterprise risk identification,real-time early warning and accurate risk control plan recommendation,and provide specific paths and plans for enterprise managers to make risk decisions.(3)Further improve the quality of enterprise risk management knowledge service.In the construction process of enterprise risk management domain knowledge graph,this dissertation uses stroke ELMo to embed IDCNN-CRF model to extract enterprise risk domain knowledge.This model not only relies on a large number of artificial features and domain knowledge,but also fully considers the internal structural characteristics of Chinese characters,and can capture the dependency information of risk domain text in a wider range to provide new methods and ideas for the construction of knowledge graph in enterprise risk management domain and further improve the quality of enterprise risk management knowledge service.Based on the knowledge graph of enterprise risk management domain,this dissertation aims to construct an intelligent enterprise risk management knowledge service model by using methods such as artificial immunity and case-based reasoning,so as to provide enterprises with knowledge services in the form of risk identification,risk early warning and risk control plan recommendation.In this study,a multidisciplinary approach is adopted to realize the efficient identification of enterprise risk,automatic early-warning analysis and intelligent decision support,and effectively solve the problem of low intelligence degree of enterprise risk management knowledge service system.However,due to the wide range of knowledge involved in the field of enterprise risk management and the lack of corpora in the field of enterprise risk management,some enterprise risk management data have single source and incomplete data acquisition.In the future,we will obtain multi-source domain data from more channels,improve the data in the field of enterprise risk management,expand the scope of knowledge in the field of enterprise risk management and further explore the enterprise risk management knowledge service model and its supporting methods.In addition,in the aspect of knowledge graph construction,this dissertation only uses text data for exploration,and does not consider multi-modal data such as picture,audio and video.In knowledge extraction in the field of enterprise risk management,due to the limited amount of manually annotated data,the accuracy rate and recall rate of experimental results need to be improved.In the future,the quality of annotated data will be improved to further enhance the effect of knowledge extraction.In terms of the construction and application of knowledge service model,this dissertation mainly takes financial service enterprises as an example to conduct empirical research,and will expand the research of enterprise types in the future.And in the link of enterprise risk quantification,this dissertation attempts to construct enterprise risk feature system from two dimensions of text vector representation and language features of core entities in the risk field,and its universality needs to be dynamically tested with more comprehensive and objective data.