Research On Operation And Maintenance Recommendation System Based On Knowledge Graph

With the development of big data and industrial Internet,data and information produced in the process of social development are everywhere.Recommending effective information to users in massive data has gradually become an important research direction.At present,the research of using recommendation technology to excavate valuable information has been very in-depth.However,the collaborative filtering algorithm,which is more representative among recommendation algorithms,faces the problems of sparsity and cold start,and also lacks the mining of data semantic information.Therefore,knowledge graph,as a semantic knowledge database describing the relationship between entities,is widely used in the field of recommendation.By studying the recommendation algorithm based on knowledge graph,this topic can realize the timely recommendation of appropriate operation and maintenance programs to employees in the field of equipment operation and maintenance.In order to improve the effect of recommendation algorithm,the knowledge graph in the field of equipment operation and maintenance is constructed and integrated with recommendation algorithm,The main research contents of this paper are as follows:(1)Firstly,the construction technology of knowledge graph is studied.By applying knowledge graph to the field of equipment operation and maintenance,the construction of knowledge graph in the field of equipment operation and maintenance is completed.After comparative analysis,Neo4j database was selected to store data,and then the domain knowledge graph of operation and maintenance was expressed vectorially.(2)Secondly,an improved recommendation algorithm based on knowledge graph is proposed.The similarity of collaborative filtering and the similarity calculation in knowledge graph are combined in a weighted way,and the fusion factor function is determined by the method of experiment traversal.The semantic similarity of knowledge graph was optimized by using weighted Euclidean distance.Then,based on the traditional scoring prediction calculation method,the time decay function is introduced to reduce the impact of the change of scoring interval on scoring prediction,so as to improve the effect of scoring prediction.Finally,the simulation results prove that the performance of the improved recommendation algorithm proposed in this paper has been improved.(3)Finally,an operation and maintenance recommendation system based on the improved collaborative filtering algorithm based on knowledge graph is designed and implemented.Through functional analysis and architecture design,Java-based SSM framework is used to realize the interaction of front-end and back-end data.Finally,the recommendation system is built,and the function of recommending the implementation scheme of abnormal operation and maintenance problems is completed,which further demonstrates the effectiveness of the algorithm.