Study And Implementation Of Knowledge Management System In Municipal Design Institute Based On Knowledge Graph

Domain knowledge and experience attained in daily operation plays a vital role in a company's development and expansion.However,problems remain unsolved in aspects like managing and utilizing those knowledge and experience.In view of the above problems,this paper designs and implements a knowledge management system of a design institute by describing the whole process of enterprise knowledge management,combing the knowledge labeling system structure of the municipal design institute,integrating the actual needs of the unit,integrating knowledge management technology and functions.Specifically,the research work of this paper mainly includes the following aspects:1.Analyze the system's business structure,construction content,functional requirements and performance requirements;then describe and analyze the business requirements,function positioning and technical architecture to transform business function requirements into system requirements.Clear key technical indicators.2.This paper presents a top-down and knowledge-graph-based approach for knowledge representation.Based on the problems of term extraction,concept extraction and hierarchy construction,this paper proposes three key algorithms.In specific,this paper presents(1)term extraction based on topic models and pretrained word vectors,(2)concept recognition based on semantic template and recursive neural network and(3)hierarchy construction based on semantic template and multi-attentional remote supervision.Moreover,this paper strikes a balance between formalization and intuitivity of the persistence of domain knowedge by using Neo4 j,the graph database.3.According to the needs analysis of the knowledge management system of the design institute,the system is functionally designed and implemented.The main class diagram design and database design and optimization clearly describe the design ideas of the knowledge management system.In specific,this paper introduce the design and implementation of knowledge classification,document center,knowledge communication and knowledge management portal.4.The key methods used in this system and the system itself are fully estimated and tested.Estimation results shows that the precision and recall of the proposed method have witnessed a considerable boost comparing to the baseline method on both term extraction and concept recognition.Test results indicate the system itself is fully functional and performance acceptable to put into use.The system has been put into operation on the Shanghai Urban Construction Design and Research Institute(Group)Co.,Ltd.The actual operation shows that the system enhances the utilization of relevant knowledge in the design institute,improves the design work efficiency,and achieves good results.