摘要
为提升大数据管理系统在不确定性数据处理与动态决策支持方面的性能,研究采用贝叶斯概率建模与分布式计算协同方法,设计包含多源数据接入、动态模型构建、任务调度、可视化决策的四层架构系统,对比分析贝叶斯算法与传统机器学习模型支持向量机(Support Vector Machine,SVM)在分类精度、响应时延、异常检测召回率等维度的性能差异。研究结果表明,该系统依托贝叶斯网络动态更新机制与资源弹性调度策略,分类准确率为89.3%,异常检测召回率为93.5%,验证了贝叶斯算法在大数据管理复杂场景下的功能价值。
To enhance the performance of big data management systems in handling uncertain data and providing dynamic decision support,the study employs a collaborative approach of bayesian probabilistic modeling and distributed computing to design a four-layer architecture system that includes multi-source data integration,dynamic model construction,task scheduling,and visual decision-making.The study compares the performance differences between bayesian algorithms and traditional machine learning models like support vector machine(SVM)in terms of classification accuracy,response latency,and anomaly detection recall rate.The results show that the system,leveraging the dynamic update mechanism of bayesian networks and resource elastic scheduling strategies,achieves a classification accuracy of 89.3%and an anomaly detection recall rate of 93.5%,validating the functional value of bayesian algorithms in complex big data management scenarios.
作者
王星语
WANG Xingyu(Shandong University of Finance and Economics,Jinan Shandong 250000,China)
出处
《信息与电脑》
2025年第18期53-55,共3页
Information & Computer
关键词
贝叶斯算法
大数据管理
分布式计算
bayesian algorithm
big data management
distributed computing
