网络安全数据分类的优化研究  

Optimisation Study of Cyber Security Data Classification

作　　者：王静奕 WANG Jingyi(Wuhan Vocational College of Communications and Publishing,Hubei Wuhan 430223,China)

机构地区：[1]武汉信息传播职业技术学院,湖北武汉430223

出　　处：《信息与电脑》2023年第13期117-119,181,共4页Information & Computer

摘　　要：为提高网络安全数据分类的准确性,对网络安全数据分类加以优化。首先通过K均值(K-means)聚类算法确定K值及选取初始中心,进行网络安全数据的聚类分析,其次以随机森林算法预处理聚类数据,消除不相关特征的数据,再次领域粗糙集算法提取数据特征,建立极端梯度提升(eXtreme Gradient Boosting,XGBoost)分类模型。结果表明,文章方法在数据分类时有较高的召回率和查准率,且所需时间较短。In order to improve the accuracy of cyber security data classification,the cyber security data classification is optimized.The K-means clustering algorithm is used to determine the K-value and select the initial centre,and then the random forest algorithm is used to pre-process the clustered data,eliminate the data with irrelevant features,and extract the data features by the domain rough set algorithm,so as to establish the XGBoost classification model.From the experimental results,the method in this paper has a high recall and accuracy rate in data classification,and takes less time.

关 键 词：网络安全 数据分类 召回率 查准率 聚类分析 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]