基于随机森林的云环境下服务故障识别方法  

Method of random forest based service fault identification in cloud environment

作　　者：邢超 姜瑛 XING Chao;JIANG Ying(Yunnan Key Laboratory of Computer Technology Application,Kunming 650500,China;School of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]云南省计算机技术应用重点实验室,云南昆明650500 [2]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《现代电子技术》2022年第16期87-92,共6页Modern Electronics Technique

基　　金：国家自然科学基金项目(61462049);国家自然科学基金项目(61063006);国家自然科学基金项目(60703116);云南省应用基础研究计划重点项目基金(2017FA033)。

摘　　要：现有服务故障识别方法大多存在准确率不高、忽略服务上下文信息等问题,为此,文中提出一种基于云环境下的服务故障识别方法。首先,使用一种综合过滤法和嵌入法的特征选择方法对服务运行数据和服务上下文数据的特征进行筛选,并使用KMeans++SMOTE算法将服务运行数据和服务上下文数据中的正常数据与故障数据平衡化;然后,使用改进加权投票决策的随机森林算法对预处理后的服务运行数据和服务上下文数据进行分类;最后,根据服务运行数据和服务上下文数据的分类结果作进一步综合判断,得到服务的故障类型。实验结果表明,所提出的服务故障识别方法可以有效提高云环境下的服务故障识别率。As most of the existing service fault identification methods have problems such as low accuracy and neglect of service context information,a service fault recognition method based on cloud environment is proposed.The feature selection method to synthesize filtering method and embedding method is used to screen the features of service operation data and the service context data,in which the KMeans++SMOTE algorithm is used to balance the normal data and fault data in the service operation data and service context data,and then the random forest algorithm with improved weighted voting decision is used to classify the pre⁃processed service operation data and service context data.According to the classification results of the service operation data and the service context data,the fault type of the service is obtained after the further comprehensive judgment.The experimental results show that the proposed service fault identification method can effectively improve the service fault identification rate in the cloud environment.

关 键 词：云计算 服务故障 故障识别 数据特征筛选 数据平衡化 随机森林 

分 类 号：TN911-34[电子电信—通信与信息系统] TP391.9[电子电信—信息与通信工程]