基于Java3D网络大数据多分辨率采集方法仿真  

Simulation of Network Big Data Multi-Resolution Acquisition Method Based on Java3D

作　　者：李晓明 应毅[1] 曾岳[2] LI Xiao-ming;YING Yi;ZENG Yue(School of Computer Science and Engineering,Sanjiang University,Nanjing Jiangsu 210012,China;School of Software Engineering,Jinling Institute of Technology,Nanjing Jiangsu 211169,China)

机构地区：[1]三江学院计算机科学与工程学院,江苏南京210012 [2]金陵科技学院软件学院,江苏南京211169

出　　处：《计算机仿真》2020年第2期416-420,共5页Computer Simulation

基　　金：江苏省高等学校自然科学研究项目(18KJB520042);江苏省高等学校自然科学研究项目(19KJB520049)。

摘　　要：针对传统的网络大数据多分辨率采集方法通常存在采集成本高、完成时间较长、采集准确率较低等问题,提出基于Java3D的大数据网络多分辨率采集方法。通过对基于Java3D网络的大数据多分辨率进行时间序列的空间重构,求出多分辨率时间序列的自相关函数,并引入平均互动数据提取方法对网络数据多分辨率采集的功率谱密度进行估计,采用ADASYN算法,对无效多分辨率数据进行去除,实现大数据多分辨率精准采集。实验结果表明,所提方法采集成本较低、完成时间较短、采集准确率较高;具有一定的实用价值,可广泛应用于各个领域。In the traditional network big data multi-resolution acquisition,the acquisition cost is high and the completion time is long.Meanwhile,the acquisition accuracy is low.Therefore,this paper proposed a multi-resolution acquisition method for network big data based on Java 3D.On the basis of the spatial reconstruction of time series,the autocorrelation function of multi-resolution time series was obtained.Moreover,the average interactive data extraction method was introduced to estimate the power spectral density of multi-resolution acquisition of network data.Finally,ADASYN algorithm was used to remove the invalid data,and thus to achieve the accurate multi-resolution acquisition of network big data.Simulation results show that the proposed method has low acquisition cost,short completion time and high acquisition accuracy.

关 键 词：大数据 多分辨率 采集 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]