摘要
为了解决网络异常数据挖掘过程中漏报率、误报率较高的问题,文章提出一种基于改进K-means聚类算法的网络异常数据挖掘与分类方法。文章通过构建并行化频繁项集挖掘环境加速数据处理,利用局部离群点检测剔除异常值,同时引入K-means聚类对数据的最大最小距离展开计算,融合隶属度函数与密度峰值优化算法,改进聚类初始中心选择及簇边界调整,从而提高异常识别准确性和分类效率。通过实验结果证明,该方法能够明显改善聚类效果与性能。
In order to solve the problem of high false positive rate and false positive rate in the process of network abnormal data mining,this paper proposes a method of network abnormal data mining and classification based on improved K-means clustering algorithm.A parallel frequent itemset mining environment is constructed to accelerate data processing,and local outlier detection is used to eliminate outliers.K-means clustering was introduced to calculate the maximum and minimum distance of the data,and membership function and density peak optimization algorithm were integrated to improve the initial center selection and cluster boundary adjustment of the cluster,so as to improve the accuracy of anomaly recognition and classification efficiency.The experimental results show that the clustering effect and performance of the proposed method are obviously improved.
作者
贺萌
HE Meng(Changzhou College of Information Technology,Changzhou 213164 China)
出处
《无线互联科技》
2024年第18期119-122,共4页
Wireless Internet Science and Technology