基于烟花算法的智慧城市物联网数据流聚类方法  

Research on Clustering Method for Smart City IoT Data Streams based on Fireworks Algorithm

作　　者：郑泛舟[1] ZHENG Fanzhou(Department of Information Engineering,Meizhouwan Vocational Technology College,Putian 351119,China)

机构地区：[1]湄洲湾职业技术学院信息工程系,福建莆田351119

出　　处：《成都工业学院学报》2024年第3期50-54,共5页Journal of Chengdu Technological University

基　　金：福建省教育厅项目(JAT220739)。

摘　　要：以往的智慧城市物联网数据流聚类方法对数据特征提取不精准,聚类速度慢。为了提高聚类速度,缩短数据流聚类的耗时,设计了基于烟花算法的智慧城市物联网数据流聚类方法。通过对样本矩阵的标准化计算,降低数据运算的压力,完成对物联网数据的预处理。在烟花算法的支持下,对数据进行去重处理,精准提取数据特征。再根据数据属性不同,计算数据的响应函数,构建数据聚类模型,利用余弦对数据进行聚类分析,实现物联网数据流的聚类。实验结果表明,该聚类方法的聚类平均耗时为15.52 ms,说明该方法能够有效缩短聚类耗时。The previous clustering methods for smart city IoT data streams are slow due to inaccurate feature extraction.In order to improve clustering speed and shorten the time required for data stream clustering,a clustering method for smart city IoT data stream based on fireworks algorithm was designed.Firstly,by standardizing the calculation of the sample matrix,the pressure of data operations was reduced,and pre-processing of IoT data was completed.Secondly,with the support of the fireworks algorithm,the data was de-duplicated to accurately extract data features.Finally,according to the different attributes of the data,the response function of the data was calculated,the data clustering model was constructed,and the cosine was used to perform clustering analysis on the data,so as to realize the clustering of IoT data streams.The experimental results show that the average clustering time of our clustering method is 15.52 ms,indicating that our method can effectively shorten the clustering time.

关 键 词：烟花算法 智慧城市 物联网 数据流聚类方法 特征提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]