自适应云资源大块数据对象并行存取方法  

Parallel Access Method for Adaptive Cloud Resource Big Data Block Object

作　　者：杨建 刘述木 YANG Jian;LIU Shu-mu(Robotics Engineering Laboratory for Sichuan Equipment Manufacturing Industry,Deyang Sichuan 618000,China;School of Software,Sichuan University,Sichuan Chengdu 610065,China)

机构地区：[1]四川省装备制造业机器人应用技术工程实验室,四川德阳618000 [2]四川大学软件学院,四川成都610065

出　　处：《计算机仿真》2021年第10期306-310,共5页Computer Simulation

摘　　要：对自适应云资源大数据块对象进行并行存取时,由于没有利用数据挖掘理论构建大数据对象的并行存取模型,导致并行存取的时间长、内存占比高、提取正确率低等问题,由此提出自适应云资源大数据块对象的并行存取方法。利用聚类算法对大数据块对象中的数据进行缺失值填充,并将填充后的数据进行分割处理,获取新的数据集;获取大数据块对象的输入层神经元,计算数据采样梯度函数,结合数据挖掘理论构建大数据块对象的并行存取模型;将自适应云资源大数据块对象放入模型中进行自适应寻优处理,以此完成自适应云资源大数据块对象的并行存取。实验结果表明,利用上述方法对大数据块对象进行并行存取时存取时间短、内存占比低、存取的正确率高。In the process of parallel access to adaptive cloud resource big data block objects, the lack of parallel access model of big data objects leads to the problems of long parallel access time, high memory proportion and low extraction accuracy. Therefore, this paper proposes a parallel access method of adaptive cloud resource big data block object. Clustering algorithm was applied to fill the missing values of large data block objects, and the filled data was segmented to obtain a new data set. The input layer neurons of big data block objects were obtained. The gradient function of data sampling was calculated. Based on the theory of data mining, the parallel access model of big block objects was established. The adaptive cloud resource big data block object was put into the model for adaptive optimization processing, completing the parallel access of adaptive cloud resource big data block object. The results show that this method has the advantages of short access time, low memory ratio and high access accuracy.

关 键 词：大数据块对象 并行存取 聚类算法 数据分割 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]