微型电脑应用2025，Vol.41Issue(4)：174-179,6.
基于贝叶斯网络的动态数据流异常值快速检测方法
A Fast Detection Method for Abnormal Values in Dynamic Data Streams Based on Bayesian Networks
陈智斌 1陆瑛 2农英雄 2黄聪 2李喆2
作者信息
- 1. 广西中烟工业有限责任公司,广西,南宁 530000||东华大学,旭日工商管理学院,上海 200000
- 2. 广西中烟工业有限责任公司,广西,南宁 530000
- 折叠
摘要
Abstract
Outliers are a common type of dynamic data streams that can reduce the quality of dynamic data flow to a certain ex-tent.To determine the abnormal state of dynamic data streams,a fast detection method for outliers in dynamic data streams based on Bayesian networks is proposed.In a continuous communication network environment,collect dynamic data streams and complete the preprocessing work of dynamic data streams from three aspects:loss data compensation,redundant data re-moval,and data standardization.Using an improved Bayesian network structure,extract dynamic data stream features,and with the support of blockchain technology,detect stable and noisy anomalies in dynamic data streams,and output statistical re-sults of dynamic data stream outliers.The conclusion drawn from performance testing experiments is that compared with tradi-tional detection methods,the optimized design method reduces the missed detection rate and false detection rate by 0.715％and 1.49％,respectively,indicating that the optimized design method has significant advantages in detection performance.关键词
贝叶斯网络/动态数据流/数据异常/快速检测Key words
Bayesian network/dynamic data streams/data anomaly/fast detection分类
计算机与自动化引用本文复制引用
陈智斌,陆瑛,农英雄,黄聪,李喆..基于贝叶斯网络的动态数据流异常值快速检测方法[J].微型电脑应用,2025,41(4):174-179,6.
