基于微分方程的大数据分类系统设计
|
摘要
基于正交分解的大数据分类系统未运用微分分类数学模型进行大数据分类,存在分类准确率低的问题。为此设计基于微分方程的大数据分类系统。该系统硬件主要包括数据采集器和存储模块,数据采集器由芯片和单片机组成,将采集的数据通过网络接口传送给网络处理器进行处理;存储模块用于储存系统中所有数据,该模块分为应用层、功能层、语义层、设计层和数据层。系统软件部分,通过建立具有二阶时滞的微分方程,及微分分类数学模型规范集约束条件,进行微分分类数学模型的构建;根据微分分类数学模型设计大数据分类代码,实现大数据分类。实验结果表明,所设计的系统大数据分类准确率高达95%,内存占用率仅为21%～32%,具有较高的分类性能。
The differential classification mathematical model is not used for big data classification in the big data classification system based on orthogonal decomposition,which causes the problem of low classification accuracy. Therefore,a big data classification system based on the differential equation is designed. The hardware of the system mainly includes a data collector and a storage module. The data collector is composed of a chip and a single chip microcomputer. The collected data is transmitted to the network processor for processing by means of the network interface. The storage module is used for storing all the data in the system,which is divided into the application layer,functional layer,semantic layer,design layer and data layer. In the software part of the system,the differential classification mathematical model is constructed by establishing the differential equation with second-order time delay and the constraint conditions for specification set of the differential classification mathematical model. The big data classification codes are designed according to the differential classification mathematical model,so as to realize big data classification. The experimental results show that the designed system has a big data classification accuracy rate of as high as 95%,a memory occupancy rate of only 21%～32%,and a high classification performance.
The differential classification mathematical model is not used for big data classification in the big data classification system based on orthogonal decomposition,which causes the problem of low classification accuracy. Therefore,a big data classification system based on the differential equation is designed. The hardware of the system mainly includes a data collector and a storage module. The data collector is composed of a chip and a single chip microcomputer. The collected data is transmitted to the network processor for processing by means of the network interface. The storage module is used for storing all the data in the system,which is divided into the application layer,functional layer,semantic layer,design layer and data layer. In the software part of the system,the differential classification mathematical model is constructed by establishing the differential equation with second-order time delay and the constraint conditions for specification set of the differential classification mathematical model. The big data classification codes are designed according to the differential classification mathematical model,so as to realize big data classification. The experimental results show that the designed system has a big data classification accuracy rate of as high as 95%,a memory occupancy rate of only 21%～32%,and a high classification performance.
引文
[1]袁芬,徐从富.基于机会协作的农业物联网大数据处理系统设计[J].计算机应用,2014,34(7):2136?2139.YUAN Fen,XU Congfu. Big data processing system based onopportunistic cooperation for agricultural Internet of Things[J].Journal of computer applications,2014,34(7):2136?2139.
[2]王永坤,罗萱,金耀辉.基于私有云和物理机的混合型大数据平台设计及实现[J].计算机工程与科学,2018,40(2):191?199.WANG Yongkun,LUO Xuan,JIN Yaohui. A hybrid big dataplatform based on private cloud VMs and bare metals[J]. Com?puter engineering&science,2018,40(2):191?199.
[3]段青玲,魏芳芳,张磊,等.基于Web数据的农业网络信息自动采集与分类系统[J].农业工程学报,2016,32(12):172?178.DUAN Qingling,WEI Fangfang,ZHANG Lei,et al. Automaticacquisition and classification system for agricultural network in?formation based on Web data[J]. Transactions of the ChineseSociety of Agricultural Engineering,2016,32(12):172?178.
[4]向小佳,赵晓芳,刘洋,等.一种正交分解大数据处理系统设计方法及实现[J].计算机研究与发展,2017,54(5):1097?1108.XIANG Xiaojia,ZHAO Xiaofang,LIU Yang,et al. A designmethod and implementation of orthogonal decomposition basedbig data processing system[J]. Journal of computer researchand development,2017,54(5):1097?1108.
[5]魏萍,丁卯,左信,等.基于微分方程对称的分布参数系统稳态控制[J].自动化学报,2014,40(10):2163?2170.WEI Ping,DING Mao,ZUO Xin,et al. Steady?state controlfor distributed parameter systems by symmetry of differentialequations[J]. Acta Automatica Sinica,2014,40(10):2163?2170.
[6]曾胜.重大危险源动态智能监测监控大数据平台框架设计[J].中国安全科学学报,2014,24(11):166?171.ZENG Sheng. Framework design of large data platform formonitoring and controlling major hazards dynamically and intel?ligently[J]. China safety science journal,2014,24(11):166?171.
[7]李雨童,姚登举,李哲,等.基于R的医学大数据挖掘系统研究[J].哈尔滨理工大学学报,2016,21(2):38?43.LI Yutong,YAO Dengju,LI Zhe,et al. Research on data min?ing system for big data in medical field based on R platform[J]. Journal of Harbin University of Science and Technology,2016,21(2):38?43.
[8]陈学斌,王师,董岩岩.面向大数据的并行分类混合算法研究[J].微电子学与计算机,2016,33(4):138?140.CHEN Xuebin,WANG Shi,DONG Yanyan. Research on par?allel classification hybrid algorithm for big data[J]. Microelec?tronics&computer,2016,33(4):138?140.
[9]胡为艳,艾民,周光彬,等.基于大数据的信令监测系统的设计与实现[J].电视技术,2016,40(1):95?101.HU Weiyan,AI Min,ZHOU Guangbin,et al. Design and im?plementation of big data based signaling monitoring system[J].Video engineering,2016,40(1):95?101.
[10]王丹,张广,陈锋,等.智能检伤分类系统的设计与研究[J].军事医学,2015,39(9):651?655.WANG Dan,ZHANG Guang,CHEN Feng,et al. Designand research of intelligent system for triage[J]. Military medi?cal sciences,2015,39(9):651?655.
[2]王永坤,罗萱,金耀辉.基于私有云和物理机的混合型大数据平台设计及实现[J].计算机工程与科学,2018,40(2):191?199.WANG Yongkun,LUO Xuan,JIN Yaohui. A hybrid big dataplatform based on private cloud VMs and bare metals[J]. Com?puter engineering&science,2018,40(2):191?199.
[3]段青玲,魏芳芳,张磊,等.基于Web数据的农业网络信息自动采集与分类系统[J].农业工程学报,2016,32(12):172?178.DUAN Qingling,WEI Fangfang,ZHANG Lei,et al. Automaticacquisition and classification system for agricultural network in?formation based on Web data[J]. Transactions of the ChineseSociety of Agricultural Engineering,2016,32(12):172?178.
[4]向小佳,赵晓芳,刘洋,等.一种正交分解大数据处理系统设计方法及实现[J].计算机研究与发展,2017,54(5):1097?1108.XIANG Xiaojia,ZHAO Xiaofang,LIU Yang,et al. A designmethod and implementation of orthogonal decomposition basedbig data processing system[J]. Journal of computer researchand development,2017,54(5):1097?1108.
[5]魏萍,丁卯,左信,等.基于微分方程对称的分布参数系统稳态控制[J].自动化学报,2014,40(10):2163?2170.WEI Ping,DING Mao,ZUO Xin,et al. Steady?state controlfor distributed parameter systems by symmetry of differentialequations[J]. Acta Automatica Sinica,2014,40(10):2163?2170.
[6]曾胜.重大危险源动态智能监测监控大数据平台框架设计[J].中国安全科学学报,2014,24(11):166?171.ZENG Sheng. Framework design of large data platform formonitoring and controlling major hazards dynamically and intel?ligently[J]. China safety science journal,2014,24(11):166?171.
[7]李雨童,姚登举,李哲,等.基于R的医学大数据挖掘系统研究[J].哈尔滨理工大学学报,2016,21(2):38?43.LI Yutong,YAO Dengju,LI Zhe,et al. Research on data min?ing system for big data in medical field based on R platform[J]. Journal of Harbin University of Science and Technology,2016,21(2):38?43.
[8]陈学斌,王师,董岩岩.面向大数据的并行分类混合算法研究[J].微电子学与计算机,2016,33(4):138?140.CHEN Xuebin,WANG Shi,DONG Yanyan. Research on par?allel classification hybrid algorithm for big data[J]. Microelec?tronics&computer,2016,33(4):138?140.
[9]胡为艳,艾民,周光彬,等.基于大数据的信令监测系统的设计与实现[J].电视技术,2016,40(1):95?101.HU Weiyan,AI Min,ZHOU Guangbin,et al. Design and im?plementation of big data based signaling monitoring system[J].Video engineering,2016,40(1):95?101.
[10]王丹,张广,陈锋,等.智能检伤分类系统的设计与研究[J].军事医学,2015,39(9):651?655.WANG Dan,ZHANG Guang,CHEN Feng,et al. Designand research of intelligent system for triage[J]. Military medi?cal sciences,2015,39(9):651?655.