摘要
为了适应大规模网络传输信息的处理与应用需求,设计基于时间序列特征提取的网络传输信息云挖掘方法。将时间序列分解模型架构至云计算平台的Spark集群上,通过并行处理方式,分解网络传输信息的时间序列。选取本体匹配方法重构分解结果的相空间,并在相空间内通过数据聚类提取网络传输信息时间序列的相似度特征并构建特征集。基于相似聚类算法,针对所构建的特征集,设置聚类优化目标函数,利用粒子群优化算法确定网络传输信息云挖掘的最优聚类中心点,从而输出云挖掘结果。实验结果表明,采用该方法挖掘网络传输数据的PBM最大值高于0.8,有效降低了云计算平台的剩余队列长度。
To meet the processing and application requirements of large⁃scale network transmission information,a network transmission information cloud mining method based on time series feature extraction is designed.Architecture the time series decomposition model onto the Spark cluster of the cloud computing platform,and decompose the time series of network transmission information through parallel processing.Select the ontology matching method to reconstruct the phase space of the decomposition results,and extract the similarity features of network transmission information time series through data clustering in the phase space and construct a feature set.Based on the similarity clustering algorithm,a clustering optimization objective function is set for the constructed feature set,and the particle swarm optimization algorithm is used to determine the optimal clustering center point for network transmission information cloud mining,thereby outputting the cloud mining results.The experimental results show that using this method to mine network transmission data has a maximum PBM value higher than 0.8,effectively reducing the remaining queue length of the cloud computing platform.
作者
葛耀武
GE Yaowu(Navy Qingdao Special Service Rehabilitation Center,Qingdao 266000,China)
出处
《电子设计工程》
2024年第12期171-175,共5页
Electronic Design Engineering
关键词
时间序列
特征提取
网络传输信息
云挖掘
相似聚类
相似度
time series
feature extraction
network transmission information
cloud mining
similarity clustering
similarity
