基于物联网技术的传感器网络样本数据定向挖掘方法  

Directional Mining Method of Sensor Network Sample Data Based on Internet of Things Technology

作　　者：陈江 CHEN Jiang(Xinjiang Industrial Vocational and Technical College,Urumqi 830022 China)

机构地区：[1]新疆工业职业技术学院,新疆乌鲁木齐830022

出　　处：《自动化技术与应用》2024年第10期104-107,143,共5页Techniques of Automation and Applications

基　　金：新疆维吾尔自治区职业教育研究项目(XJZJKT-2021Y19)。

摘　　要：为提升传感器网络样本数据定向挖掘效果,提出基于物联网技术的传感器网络样本数据定向挖掘方法。设计基于物联网技术的数据处理平台,包括物联网节点、边缘数据汇聚设备和DAE-RNI模型。利用DAE-RNI模型优化基于马氏距离的不完整数据填充算法,实现传感器网络样本不完整数据的填充。利用XGBoost模型定向挖掘填充后的数据。测试结果表明,设计方法定向挖掘数据的均方根误差低于0.5%,平均绝对误差低于1.0%。To improve the effectiveness of sensor network sample data directional mining,this paper proposes a sensor network sample data directional mining method based on Internet of Things technology.It designs a data processing platform based on IoT technology,including IoT nodes,edge data aggregation devices,and DAE-RNI models.The DAE-RNI model is used to optimize the incom-plete data filling algorithm based on Mahalanobis distance to fill the incomplete data of sensor network samples.It utilizes the XGBoost model to directionally mine the filled data.The test results show that the root mean square error of the directional min-ing data of the design method is less than 0.5%,and the average absolute error is less than 1.0%.

关 键 词：物联网 DAE-RNI模型 传感器 XGBoost模型 决策树 数据挖掘 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]