随机森林在气象网络入侵检测中的应用  

Research on Intrusion Detection for Meteorological Networks Based on Random Forest Method

作　　者：赵忠凯 王祝先[1] 韩书新 董洋[1] 关兴民[1] ZHAO Zhong-kai;WANG Zhu-xian;HAN Shu-xin;DONG Yang;GUAN Xing-min(Heilongjiang Provincial Meteorological Data Center,Harbin 150030 China;Heilongjiang Provincial Ecological Meteorological Center,Harbin 150030 China)

机构地区：[1]黑龙江省气象数据中心,黑龙江哈尔滨150030 [2]黑龙江省生态气象中心,黑龙江哈尔滨150030

出　　处：《自动化技术与应用》2024年第7期129-133,共5页Techniques of Automation and Applications

摘　　要：随着气象业务信息化程度的加深,气象网络在数据传输、处理和共享方面的作用越来越重要。然而,这也使得气象系统网络面临更多的安全威胁。近些年机器学习和人工智能技术迅猛发展,为气象信息系统网络安全提供了新的解决方案。为此提出基于随机森林算法的气象网络的入侵检测模型,并借助仿真实验验证模型的有效性。在仿真实验中讨论数据的预处理方法包括类别特征的独热编码,归一化以及样本不平衡的重采样技术。仿真实验表明随机森林在网络入侵检测准确率和效率都要高于K近邻算法和决策树算法,为气象平台网络安全防护提供一种新思路和新模式。With the informatization of meteorological systems,meteorological networks becomes more and more important in data transmission,processing,and sharing.However,this also poses more security threats to meteorological networks.In recent years,the rapid development of machine learning and artificial intelligence technology provids new solutions for the security of meteorological information systems.This article introduces the security protection technology of meteorological networks,uses random forest algorithm to implement the intrusion detection model,and verifies the effectiveness of the model with simulation experiment.In the simulation experiment,the paper discusses the data preprocessing methods including one-hot encoding of category features,normalization,and resampling techniques.In the experiment,it conductes comparative experiments using K-nearest neighbor,decision tree,and random forest.The simulation experiment shows that the random forest algorithm has higher accuracy and efficiency in network intrusion detection than other algorithms.This paper provids a new idea and new mode for the security protection of meteorological platform networks.

关 键 词：入侵检测 气象网络安全 随机森林 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]