首页|基于滑动聚类的窄带物联网特征级异构数据融合方法

基于滑动聚类的窄带物联网特征级异构数据融合方法

扫码查看
由于物联网数据具有多样化特征,不同类数据需求不明确,导致数据特征集提取难度较大.为了提高物联网异构数据融合效果,提出一种基于滑动聚类的窄带物联网特征级异构数据融合方法.利用平移变换得到网络数据的时间信息,通过小波变换增强无线传感器节点采集的异构数据质量;计算滑动窗差值,确定物联网特征级异构数据的初始聚类点,利用均值漂移算法完成异构数据状态聚类;通过凝聚机制完成相似数据的特征集提取,实现特征级异构数据的融合.实验结果证明,所提方法有效降低了物联网数据规模,减少了数据融合误差,且融合后信息不失真,在数据处理领域具有较高的应用价值.
Feature-level Heterogeneous Data Fusion Method for Narrowband Internet of Things Based on Sliding Clustering
Due to the diverse characteristics of Internet of Things data and unclear requirements for different types of data,it is difficult to extract data feature sets.To improve the effect of heterogeneous data fusion in the Internet of Things,a feature lev-el heterogeneous data fusion method for narrowband Internet of Things based on sliding clustering is proposed.The temporal information of network data is obtained through translation transformation,and the quality of heterogeneous data collected by wireless sensor nodes is enhanced through wavelet transform.The sliding window difference value is caculated to determine the initial clustering point of the feature-level heterogeneous data in the Internet of Things,and the mean shift algorithm is used to complete the clustering of heterogeneous data states.The feature set extraction of similar data is completed through the aggre-gation mechanism to achieve the fusion of feature-level heterogeneous data.Experimental results show that the proposed meth-od effectively reduces the scale of Internet of Things data,diminish data fusion errors,and the information is not distorted after fusion,which has high application value in the field of data processing.

heterogeneous data in the Internet of Thingssliding windowdata fusionwavelet transformdata sihouette coef-ficientcluster center

郝亚平

展开 >

常州工业职业技术学院,信息化中心,江苏,常州 213164

物联网异构数据 滑动窗 数据融合 小波变换 数据轮廓系数 聚类中心

2024

微型电脑应用
上海市微型电脑应用学会

微型电脑应用

CSTPCD
影响因子:0.359
ISSN:1007-757X
年,卷(期):2024.40(7)