基于微积分分类数学模型的关联挖掘改进方法

详细信息    查看全文 | 推荐本文 |

英文篇名：Improved method of correlation mining based on mathematical model of calculus classification 作者：罗红英 英文作者：LUO Hongying;College of Mathematics and Statistics,Qujing Normal University; 关键词：关联挖掘 ; 并行算法改进 ; 微积分分类 ; 奇异值分解 ; 大数据挖掘 ; 数学模型 英文关键词：correlation mining;;parallel algorithm improvement;;calculus classification;;singular value decomposition;;big data mining;;mathematical model 中文刊名：XDDJ 英文刊名：Modern Electronics Technique 机构：曲靖师范学院数学与统计学院; 出版日期：2019-02-21 11:13 出版单位：现代电子技术 年：2019 期：v.42;No.535 基金：国家自然科学基金项目(11361048);; 云南省应用基础研究基金项目(2017FH001-014);; 曲靖师范学院重点课程建设项目(ZDKC2016002)~~ 语种：中文; 页：XDDJ201908030 页数：5 CN：08 ISSN：61-1224/TN 分类号：143-147

摘要

基于弱关联挖掘模型的关联挖掘方法,依据大数据流信息间的局部关联实现数据挖掘,未考虑数据信息流间的互信息特征,挖掘效果差。研究大数据关联挖掘的改进方法,采用微积分分类数学模型改进关联挖掘过程。该方法提取混合云环境下数据信息流的互信息特征,依据该特征采集大数据流模型的最大Lyapunove指数谱特征,通过矩阵压缩方法使得高维矩阵转换成低维矩阵。在此基础上依据微积分极值原理构建大数据的微积分分类数学模型,该模型通过最大Lyapu-nove指数谱网格分布矩阵的奇异值分解方法,分解大数据特征向量矩阵行,将大数据关联挖掘过程转换成小规模并行运算过程,实现大数据挖掘中并行算法的改进。实验结果表明,采用该方法进行关联挖掘运算的时间开销的平均值为4.7 s,扩展率平均为0.7,挖掘效果佳。
        An improved big data correlation mining method is studied. The calculus classification mathematical model is adopted to improve the correlation mining process. In this method,the mutual information characteristics of data information flows in the mixed cloud environment are extracted. The maximum Lyapunove exponential spectrum characteristics of the big data flow model are collected according to the extracted characteristics. The high-dimensional matrix is converted to the low-dimensional matrix by means of the matrix compression method. On this basis,the calculus classification mathematical model of big data is constructed according to the calculus extremum principle. In the model,the singular value decomposition method based on the grid distribution matrix of maximum Lyapunove exponential spectrums is used to decompose the matrix rows of big data eigenvectors. The big data correlation mining process is converted into the small-scale parallel operation process,so as to realize the improvement of parallel algorithms in big data mining. The experimental results show that the proposed method for correlation mining operation has an average overhead time of 4.7 s,and an average expansion rate of 0.7,which has a good mining effect.

引文

[1]李琴,傅莺莺,莫立坡.浅谈微积分所蕴含的特征值计算思想方法[J].数学的实践与认识,2016,46(8):292-296.LI Qin, FU Yingying, MO Lipo. Note on mathematical thoughts implied in calculus for solving eigenvalue problems[J]. Mathematics in practice and theory,2016,46(8):292-296.
    [2]廉文武,傅凌玲,黄潮.云计算环境下数据弱关联挖掘模型的仿真[J].计算机仿真,2015,32(4):359-362.LIAN Wenwu, FU Lingling, HUANG Chao. Simulation on weak association mining model of data in cloud computing envi-ronment[J]. Computer simulation,2015,32(4):359-362.
    [3]潘大胜,陈志福,覃焕昌.基于模糊关联迭代分区的挖掘优化方法研究[J].科学技术与工程,2016,16(24):235-238.PAN Dasheng,CHEN Zhifu,QIN Huanchang. Based on the fuzzy correlation iterative partition optimization of mining meth-od study[J]. Science technology and engineering,2016,16(24):235-238.
    [4]陈泽华,张裕,谢刚.基于粒计算的最简决策规则挖掘算法[J].控制与决策,2015,30(1):143-148.CHEN Zehua,ZHANG Yu,XIE Gang. Mining algorithm for concise decision rules based on granular computing[J]. Control and decision,2015,30(1):143-148.
    [5]顾苏杭,王士同.基于数据点本身及其位置关系辅助信息挖掘的分类方法[J].模式识别与人工智能,2018,31(3):197-207.GU Suhang,WANG Shitong. Classification approach by min-ing betweenness information beyond data points themselves[J].Pattern recognition and artificial intelligence,2018,31(3):197-207.
    [6]曹伟,龙华,杜庆治,等.基于网络微积分的QoS接入策略[J].云南大学学报(自然科学版)2016,38(1):29-36.CAO Wei,LONG Hua,DU Qingzhi,et al. Study on QoS ac-cess strategy based on network calculus[J]. Journal of Yunnan University(Natural sciences edition),2016,38(1):29-36.
    [7]郭瑞凯,丁建华,赵奎,等.充填体的分数阶微积分蠕变本构模型及其在FLAC3D中的开发应用[J].中国钨业,2017,32(5):27-31.GUO Ruikai,DING Jianhua,ZHAO Kui,et al. Fractional cal-culus creep constitutive model of filling body and its develop-ment and application in FLAC[J]. China tungsten industry,2017,32(5):27-31.
    [8]杨柳华,王洪江,吴爱祥,等.基于循环系统的膏体浓密机底流调控及其数学模型[J].工程科学学报,2017,39(10):1507-1511.YANG Liuhua,WANG Hongjiang,WU Aixiang,et al. Regula-tion and a mathematical model of underflow in paste thickeners based on a circular system design[J]. Chinese journal of engi-neering,2017,39(10):1507-1511.
    [9] ZHANG H,JI P,WANG J,et al. An improved weighted corre-lation coefficient based on integrated weight for interval neutro-sophic sets and its application in multi-criteria decision-making problems[J]. International journal of computational intelligence systems,2015,8(6):1027-1043.
    [10]徐春,李广原,王玄,等.一种基于倒排索引树的增量更新关联挖掘算法[J].计算机工程与科学,2016,38(5):1039-1045.XU Chun,LI Guangyuan,WANG Xuan,et al. An incremen-tal updating association rule mining algorithm based on invert-ed index tree[J]. Computer engineering&science,2016,38(5):1039-1045.