摘要
随着移动通信网络的发展,未来网络逐渐趋于异构化、密集化,如何对网络故障进行高效的诊断与分析面临着巨大挑战。传统基于数据的网络故障诊断方法存在可解释性差、应用性低等问题,结合知识图谱技术,提出了一种基于知识和数据双驱动的网络故障分析方法。首先通过本体构建、知识抽取以及知识融合等步骤利用Neo4j图数据库搭建面向网络故障诊断的知识图谱;然后结合机器学习进行智能化网络故障诊断与分析,将网络故障诊断问题拆分成不同子问题,对比不同机器学习算法的准确性,为不同诊断问题匹配准确度最高的机器学习算法;并利用Neo4j图数据库提出基于子图匹配的知识检索方法,将网络分析结果以知识图谱子图的形式展示。仿真结果表明,所提方法可以有效提高网络故障诊断的准确性,提高了在实际工程中的应用性。
With the development of mobile communication networks,future networks tend to be increasingly heterogeneous and dense.So how to diagnose and analyze network faults efficiently is facing great challenges.Traditional network fault diagnosis methods based on data have problems of poor interpretability and low applicability.In this paper,a network fault analysis method based on knowledge and data was proposed by combining knowledge graph technology.Firstly,the knowledge graph for network fault diagnosis was constructed by using Neo4j graph database through ontology construction,knowledge extraction and knowledge fusion.Then the intelligent network fault diagnosis and analysis was carried out by combining machine learning.The network fault diagnosis problems were divided into different sub-problems,and the accuracy of different machine learning algorithms was compared to match the machine learning algorithm with the highest accuracy for different diagnosis problems.A knowledge retrieval method based on subgraph matching was proposed by using Neo4j graph database,and the results of network analysis were presented in the form of knowledge graph subgraph.Simulation results showed that the proposed method can effectively improve the accuracy of network fault diagnosis,and the application of the proposed method in practical engineering was improved by displaying the fault analysis results in the form of knowledge graph.
作者
谷奉锦
贺楚闳
潘庆亚
王晔
朱晓荣
GU Fengjin;HE Chuhong;PAN Qingya;WANG Ye;ZHU Xiaorong(School of Communication and Information Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;China Mobile Jiangsu Co.,Ltd.,Nanjing 210003,China)
出处
《无线电通信技术》
2022年第4期751-757,共7页
Radio Communications Technology
基金
国家自然科学基金(61871237,92067101)
江苏省重点研发计划(BE2021013-3)。
关键词
知识图谱
故障诊断
机器学习
knowledge graph
fault diagnosis
machine learning