粗集—支持向量机方法的软测量应用研究
|
摘要
软测量技术为解决工业控制系统在线检测过程中,一些控制参数无法通过检测仪表进行直接检测或由于检测仪表价格昂贵难以应用等实际难题提供了有效技术手段。
基于统计学习理论的支持向量机是一种新型的学习方法,它遵照结构风险最小化原则,同时支持向量机算法是一个凸二次优化问题,能够保证找到的极值解就是全局最优解,从而在统计样本量较少的情况下获得良好的统计规律和更好的泛化能力,为解决小样本、非线性、高维数等学习问题提供了一个框架,帮助解决了许多其他学习方法难以解决的问题。
粗糙集理论是一种研究不完整、不确定知识和数据的表达、学习、归纳的理论方法。其在不需额外的先验知识的情况下,通过描绘知识表达中不同属性的重要性,进行知识表达空间约简,去掉冗余信息,简化信息的表达空间维数。
由于实际生产中获得的数据往往是有限的,并且一个复杂的过程控制系统中,往往需要大量传感器监测过程参数,传感器的数量直接影响系统的投资。如何减少一些测量不重要参数的传感器,在有限的小样本基础上保证有较高的预测精度,这是本文的重点。
文章在分析了目前常用的几种软测量建模方法的基础上,结合支持向量机和粗糙集的各自特点,提出了一种基于粗集-支持向量机的软测量建模方法。利用粗糙集属性约简方法减少属性数;利用支持向量机解决小样本、非线性系统的建模问题。
本文沿着“软测量技术粗糙集理论支持向量机理论粗糙集-支持向量机混合方法的提出将粗糙集-支持向量机混合方法应用到软测量中”的思路对作者在课题研究中所取得的成果进行介绍。其中粗糙集-支持向量机混合方法和将该混合方法应用到软测量建模是本论文的重点。在课题研究过程中,除了分析分析说明了该混合方法的特点、使用步骤以及为何将它应用到软测量建模中外,主要使用工具软件ROSETTA、LIBSVM以及MATLAB分别对粗糙集理论属性约简的算法和支持向量机分类、回归建模方法进行了模拟仿真,并以某市污水处理厂的实测数据,针对污水处理过程控制参数及水质参数进行了基于粗集-支持向量机混合方法的软测量建模研究,对污水处理重要参数COD、TN、TP以及SVI等难以在线测量的参数或虽能在线测量但检测费用高的参数进行了测量,将获得的预测结
Soft Sensor Technique is an effective approach which is used to solve some actual puzzles, caused by the unavailability of measurers in the controlling process variables measuring or the high price of measurers, in industrial process control systems.
Support Vector Machine (SVM) is a novel powerful machine learning method based on statistical learning theory which observes the principle of structural risk minimization (ERM). Furthermore, the SVM algorithm is a quadratic programming problem that promises the extremal solution is the global optimum and thus makes it possible to have good learning and generalization performance under the situation of small-samples. SVM provides a framework to solve the learning problems characterized by small samples, nonlinearity, high dimension and local minima, which are difficult to be dealt with by other learning methods.
Rough set theory is a theoretical method which is applied to study the expression, learning and inducing of incomplete and uncertain data. It can reduce the knowledge- expression space, cancel the redundant information via describing the importance of different attribute in knowledge expression without prior knowledge.
Since data acquired in practical production are always limited and a great number of sensors are demanded in a complicated process control system, the number of sensors has a direct influence on the investment. The key point of this thesis is how to decrease some sensors testing unimportant variables and insures a high predictive classification accuracy on the bases of limited small-samples.
This thesis proposes a soft-sensor modeling method based on RS-SVM after analyzing some common soft-sensor modeling methods and considering the advantage of SVM and RS. refining the number of properties by using the attribute reduction theory of RS and taking advantage of the good generalization performance of SVM to modeling for the Soft-sensor problem of small examples, nonlinear and high dimensions.
This dissertation is along the thinking of " Soft Sensor Technique Rough Set Theory Support Vector Machine Theory the combination method of RS and SVM the application of RS-SVM method in Soft-Sensor " to introduce the achievement concluded in the subject study. The most essential part is the last two parts, namely, the combination method of RS and SVM, as well as the application of RS-SVM method in
基于统计学习理论的支持向量机是一种新型的学习方法,它遵照结构风险最小化原则,同时支持向量机算法是一个凸二次优化问题,能够保证找到的极值解就是全局最优解,从而在统计样本量较少的情况下获得良好的统计规律和更好的泛化能力,为解决小样本、非线性、高维数等学习问题提供了一个框架,帮助解决了许多其他学习方法难以解决的问题。
粗糙集理论是一种研究不完整、不确定知识和数据的表达、学习、归纳的理论方法。其在不需额外的先验知识的情况下,通过描绘知识表达中不同属性的重要性,进行知识表达空间约简,去掉冗余信息,简化信息的表达空间维数。
由于实际生产中获得的数据往往是有限的,并且一个复杂的过程控制系统中,往往需要大量传感器监测过程参数,传感器的数量直接影响系统的投资。如何减少一些测量不重要参数的传感器,在有限的小样本基础上保证有较高的预测精度,这是本文的重点。
文章在分析了目前常用的几种软测量建模方法的基础上,结合支持向量机和粗糙集的各自特点,提出了一种基于粗集-支持向量机的软测量建模方法。利用粗糙集属性约简方法减少属性数;利用支持向量机解决小样本、非线性系统的建模问题。
本文沿着“软测量技术粗糙集理论支持向量机理论粗糙集-支持向量机混合方法的提出将粗糙集-支持向量机混合方法应用到软测量中”的思路对作者在课题研究中所取得的成果进行介绍。其中粗糙集-支持向量机混合方法和将该混合方法应用到软测量建模是本论文的重点。在课题研究过程中,除了分析分析说明了该混合方法的特点、使用步骤以及为何将它应用到软测量建模中外,主要使用工具软件ROSETTA、LIBSVM以及MATLAB分别对粗糙集理论属性约简的算法和支持向量机分类、回归建模方法进行了模拟仿真,并以某市污水处理厂的实测数据,针对污水处理过程控制参数及水质参数进行了基于粗集-支持向量机混合方法的软测量建模研究,对污水处理重要参数COD、TN、TP以及SVI等难以在线测量的参数或虽能在线测量但检测费用高的参数进行了测量,将获得的预测结
Soft Sensor Technique is an effective approach which is used to solve some actual puzzles, caused by the unavailability of measurers in the controlling process variables measuring or the high price of measurers, in industrial process control systems.
Support Vector Machine (SVM) is a novel powerful machine learning method based on statistical learning theory which observes the principle of structural risk minimization (ERM). Furthermore, the SVM algorithm is a quadratic programming problem that promises the extremal solution is the global optimum and thus makes it possible to have good learning and generalization performance under the situation of small-samples. SVM provides a framework to solve the learning problems characterized by small samples, nonlinearity, high dimension and local minima, which are difficult to be dealt with by other learning methods.
Rough set theory is a theoretical method which is applied to study the expression, learning and inducing of incomplete and uncertain data. It can reduce the knowledge- expression space, cancel the redundant information via describing the importance of different attribute in knowledge expression without prior knowledge.
Since data acquired in practical production are always limited and a great number of sensors are demanded in a complicated process control system, the number of sensors has a direct influence on the investment. The key point of this thesis is how to decrease some sensors testing unimportant variables and insures a high predictive classification accuracy on the bases of limited small-samples.
This thesis proposes a soft-sensor modeling method based on RS-SVM after analyzing some common soft-sensor modeling methods and considering the advantage of SVM and RS. refining the number of properties by using the attribute reduction theory of RS and taking advantage of the good generalization performance of SVM to modeling for the Soft-sensor problem of small examples, nonlinear and high dimensions.
This dissertation is along the thinking of " Soft Sensor Technique Rough Set Theory Support Vector Machine Theory the combination method of RS and SVM the application of RS-SVM method in Soft-Sensor " to introduce the achievement concluded in the subject study. The most essential part is the last two parts, namely, the combination method of RS and SVM, as well as the application of RS-SVM method in
引文
[1] 荣冈,金晓明,王树青.先进控制技术应用.第三讲.软测量技术及其应用.化工自动化及仪表,1999,26(4):70～74.
[2] 骆晨钟,邵慧鹤.软仪表技术及其应用[J].仪表技术与传感器,1999,(1):32～39.
[3] 孙欣,王金春,何声亮.过程软测量[J].自动化仪表.1995,16(8):1~5.
[4] 王士同,神经模糊系统及其应用[M].北京:北京航空航天大学出版社,1998.45～56.
[5] 周鸣争.基于神经网络的软测量模型应用[J].仪器仪表学报,1999,20(6):654～656.
[6] 于静江,周春晖.过程控制中的软测量技术[J].控制理论与应用,1996,13(4):137～144.
[7] 俞金寿,刘爱伦.软测量技术及其应用[J].世界仪表和自动化,1997,1(2):18～20.
[8] 徐敏,俞金寿.软测量技术[J].石油化工自动化,1998,(2):1～3.
[9] 朱学峰.软测量技术及其应用[J].华南理工大学学报(自然科学版),2002,30(11):61～67.
[10] 骆晨钟,邵惠鹤,吴俊生.基于机理分析的方法实现催化裂化反应预测转化率、产品产率的软测量[J].化工自动化及仪表,1999,26(2):29～33.
[11] Adilson Jose de Assis, Rusbens Maciel Filho.Soft Sensors development for on-line bioreactor state estimation[J].Computer and Chemical Engineering,2000,24:1009～1103.
[12] 冯丽辉,毕贵红,赖华,彭向华.A2/O 污水处理过程出水水质的多元线性回归软测量模型[J].测控技术,2003,22(6):11～14.
[13] 佘颍,陈五平.喷射塔中 SO2 吸收传到系数的关联合回归[J].化学工程,1990,18(2):67～69.
[14] 陈常贵,姚玉英.筛选精馏塔点效率的经验关联[J].化学工程,1993,21(5):11～14.
[15] Pawlak Z.Rough set approach to Knowledge-based decision support [J].European Journal of Op-erational Research,1997.
[16] Golan R and Ziarko W. Methodology for stockmarket analysis utilizing rough set theory [J].Proc. Of IEEE/IAFE Conference on Computational intelligence for Financial Engineering, New Jersey, 1995.
[17] Nejman D .A rough set based method of handwritten numberals classification [J]. Institute of Computer Science Reports, Warsaw University of Technology, Warsaw,1994.
[18] Hu Xiaohua et al. Mining Knowledge rules from database-a rough set approach [J]. Proco. of IEEE International Conference on Data Engineering, Los Alamitos, 1996.
[19] Tsumoto Sh et al. Extraction of domain knowledge from database on rough set theory [J].IEEE international Conference on Fuzzy Systems, New Jersey,1996.
[20] Jelonek J et al. Rough set reduction of attributes and their domains for neural networks [J]. Computational Intelligence,1995.
[21] 邓乃杨,田英杰.数据挖掘中的新方法-支持向量机[M]。科学出版社,2004.6
[22] 马勇,黄德先.基于支持向量机的软测量建模方法[J].信息与控制,2004,33(4):417～421
[23] Cortes,C.,Vapnik,V.Support Vector networks[J].Machine Learning, 1995, 70(1):1～25.
[24] Gunn S R.Support Vector Machines for Classification and Regression[R]. UK:Universtiy of Southampton,1997.
[25] Buerges CJC. A Turtorial on Support Vector Machines for Pattern Recognition[J].Data Mining and Knowledge Discovery,1998,2(2):121~167。
[26] Osuna E, Freund R, Girosi F.Training Support Vector Machines:An Application to Face Detection[A].Proceeding soft IEEE Conference on CVPR’97[C].Puerto Rico,1997.130~136.
[27] Suykens JAK. Optimal Control By Least Squares Support Vector Machines[J].Neural Networks,2001,14(1):23～25.
[28] 李晓黎,刘继敏,史忠值.基于支持向量机与无监督聚类相结合的中文网页分类器[J].计算机学报,2001,24(1):62-68
[29] 王晓丹,王积勤.支持向量机研究与应用[J].空军工程大学学报(自然科学版).2004,5(3): 49～55
[30] M.H?ck,M.k?hne. Estimation of wastewater process parameters using neural networks[J]. water. sci. technol, 1996,33(1): 101~115
[31] Dong-jin Choi,Heekyung Park.A hybrid artificial neural network as software sensor for optimal control of a wastewater treatment process[J].water.res,2001,35(16):3959~3967
[32] Dae SungLee,Jong Moon Park. Neural network modeling for on-line estimation of nutrient dynamics in a sequentially-operated batch reactor[J].J.Biotechnology, 1999,75:229~239
[33] Alcaraz-Gonzálea, J.Harmand, A.rapaport,J. P.Steryer, V.Gonzálea- Alvarez, C.pelayo-Ortiz. Software sensors for highly uncertain WWTPs-a new approach based on interval observers. water.res. 2002,36: 2515～2524.
[34] 任敏,王万良,李探微等.基于神经网络元网络的污水处理软测量系统的研究.自动化仪表[J],2001,22(10):8～9
[35] 冯丽辉,毕贵红,赖华等 A2/O 污水处理过程出水水质的多元线性回归软测量模型[J]。测控技术,2003,22(6):11～14
[36] 陈守,陈小冰.模糊模式识别模型及其在化工中的应用[J].化工学报,1991,(6):660～668.
[37] 罗荣富.软测量方法及其工业应用[A].第六届过程控制科学报告会论文集[C].上海:上海交通大学出版社,1993.324-329.
[38] Pawlak Z . Rough set approach to Knowledge-based decision support [J].European Journal of Op-erational Research,1997.
[39] Skowron.A.The discernibility matrices and functinns in Information system. R.. Slowinski. Intelligent decision support.Handbook of Applications and Advances of the rough set theory[J]. Kluwer Academic Publishers.1992,331~363
[40] 张文修等,编著.粗糙集理论与方法[M].北京:科学出版社,2001
[41] Tengfei Zhang,Jianmei Xiao.Dynamic System Modeling based on Rough Sets and RBF Neural Networks.Proceedings of the 5th World Congress on Intelligent Control and Automation[J], June 15-19,2004,Hangzhou,P.R.China,pp.185-189
[42] 王东锴,梁樑．含冲突的决策表中的决策规则发现[J].系统工程,2003.vol.21,No.l
[43] V.N.Vapnik.The Nature of Statistical Learning Theory[J].SPringer New York,1995
[44] 张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(l):32 一 42
[45] C.Burges.A tutorial on support vector machines for pattern recognition.Data Mining and Knowledge Discovery,(2)1998:121~ 167
[46] 边肇祺,张学工.模式识别[M],清华大学出版社,北京,2000
[47] 余建平.基于改进神经网络的污水处理参数软测量模型研究[D].重庆:重庆大学城市建设与环境工程学院,2005
[48] 卿晓霞,余建平,王波.基于粗糙集-神经网络的污水参数软测量技术研究[J],2005.3.
[49] 王国胤.Rough 集理论与知识获取[M],西安交通大学出版社.
[50] 张自杰,林荣忱,金儒霖.排水工程[M].第四版.北京:中国建筑工业出版社,2006.6,103～107.
[51] 奚旦立,孙裕生,刘秀英.环境检测[M].第二版.北京:高等教育出版社,2000,427～465
[52] Chih-Wei Hsu, Chih-Chung Chang, and Chih-Jen Lin. A Practical Guide to Support Vector Classification.http://www.csite.ntu.edu.tw.
[53] Hsuan-Tien Lin and Chih-Jen Lin. A Study on Sigmoid Kernels for SVM and Training of non-PSD Kernels by SMO-type Methods.http://www.csie.ntu.edu.tw/~cjlin/papers/tanh.pdf
[54] 李海青,黄志尧等.软测量技术原理及应用[M].化学工业出版社.2000.9
[55] 张文艺,钟梅英等.活性污泥法系统人工神经网络模型.给水排水,2002,28(6):12~15
[2] 骆晨钟,邵慧鹤.软仪表技术及其应用[J].仪表技术与传感器,1999,(1):32～39.
[3] 孙欣,王金春,何声亮.过程软测量[J].自动化仪表.1995,16(8):1~5.
[4] 王士同,神经模糊系统及其应用[M].北京:北京航空航天大学出版社,1998.45～56.
[5] 周鸣争.基于神经网络的软测量模型应用[J].仪器仪表学报,1999,20(6):654～656.
[6] 于静江,周春晖.过程控制中的软测量技术[J].控制理论与应用,1996,13(4):137～144.
[7] 俞金寿,刘爱伦.软测量技术及其应用[J].世界仪表和自动化,1997,1(2):18～20.
[8] 徐敏,俞金寿.软测量技术[J].石油化工自动化,1998,(2):1～3.
[9] 朱学峰.软测量技术及其应用[J].华南理工大学学报(自然科学版),2002,30(11):61～67.
[10] 骆晨钟,邵惠鹤,吴俊生.基于机理分析的方法实现催化裂化反应预测转化率、产品产率的软测量[J].化工自动化及仪表,1999,26(2):29～33.
[11] Adilson Jose de Assis, Rusbens Maciel Filho.Soft Sensors development for on-line bioreactor state estimation[J].Computer and Chemical Engineering,2000,24:1009～1103.
[12] 冯丽辉,毕贵红,赖华,彭向华.A2/O 污水处理过程出水水质的多元线性回归软测量模型[J].测控技术,2003,22(6):11～14.
[13] 佘颍,陈五平.喷射塔中 SO2 吸收传到系数的关联合回归[J].化学工程,1990,18(2):67～69.
[14] 陈常贵,姚玉英.筛选精馏塔点效率的经验关联[J].化学工程,1993,21(5):11～14.
[15] Pawlak Z.Rough set approach to Knowledge-based decision support [J].European Journal of Op-erational Research,1997.
[16] Golan R and Ziarko W. Methodology for stockmarket analysis utilizing rough set theory [J].Proc. Of IEEE/IAFE Conference on Computational intelligence for Financial Engineering, New Jersey, 1995.
[17] Nejman D .A rough set based method of handwritten numberals classification [J]. Institute of Computer Science Reports, Warsaw University of Technology, Warsaw,1994.
[18] Hu Xiaohua et al. Mining Knowledge rules from database-a rough set approach [J]. Proco. of IEEE International Conference on Data Engineering, Los Alamitos, 1996.
[19] Tsumoto Sh et al. Extraction of domain knowledge from database on rough set theory [J].IEEE international Conference on Fuzzy Systems, New Jersey,1996.
[20] Jelonek J et al. Rough set reduction of attributes and their domains for neural networks [J]. Computational Intelligence,1995.
[21] 邓乃杨,田英杰.数据挖掘中的新方法-支持向量机[M]。科学出版社,2004.6
[22] 马勇,黄德先.基于支持向量机的软测量建模方法[J].信息与控制,2004,33(4):417～421
[23] Cortes,C.,Vapnik,V.Support Vector networks[J].Machine Learning, 1995, 70(1):1～25.
[24] Gunn S R.Support Vector Machines for Classification and Regression[R]. UK:Universtiy of Southampton,1997.
[25] Buerges CJC. A Turtorial on Support Vector Machines for Pattern Recognition[J].Data Mining and Knowledge Discovery,1998,2(2):121~167。
[26] Osuna E, Freund R, Girosi F.Training Support Vector Machines:An Application to Face Detection[A].Proceeding soft IEEE Conference on CVPR’97[C].Puerto Rico,1997.130~136.
[27] Suykens JAK. Optimal Control By Least Squares Support Vector Machines[J].Neural Networks,2001,14(1):23～25.
[28] 李晓黎,刘继敏,史忠值.基于支持向量机与无监督聚类相结合的中文网页分类器[J].计算机学报,2001,24(1):62-68
[29] 王晓丹,王积勤.支持向量机研究与应用[J].空军工程大学学报(自然科学版).2004,5(3): 49～55
[30] M.H?ck,M.k?hne. Estimation of wastewater process parameters using neural networks[J]. water. sci. technol, 1996,33(1): 101~115
[31] Dong-jin Choi,Heekyung Park.A hybrid artificial neural network as software sensor for optimal control of a wastewater treatment process[J].water.res,2001,35(16):3959~3967
[32] Dae SungLee,Jong Moon Park. Neural network modeling for on-line estimation of nutrient dynamics in a sequentially-operated batch reactor[J].J.Biotechnology, 1999,75:229~239
[33] Alcaraz-Gonzálea, J.Harmand, A.rapaport,J. P.Steryer, V.Gonzálea- Alvarez, C.pelayo-Ortiz. Software sensors for highly uncertain WWTPs-a new approach based on interval observers. water.res. 2002,36: 2515～2524.
[34] 任敏,王万良,李探微等.基于神经网络元网络的污水处理软测量系统的研究.自动化仪表[J],2001,22(10):8～9
[35] 冯丽辉,毕贵红,赖华等 A2/O 污水处理过程出水水质的多元线性回归软测量模型[J]。测控技术,2003,22(6):11～14
[36] 陈守,陈小冰.模糊模式识别模型及其在化工中的应用[J].化工学报,1991,(6):660～668.
[37] 罗荣富.软测量方法及其工业应用[A].第六届过程控制科学报告会论文集[C].上海:上海交通大学出版社,1993.324-329.
[38] Pawlak Z . Rough set approach to Knowledge-based decision support [J].European Journal of Op-erational Research,1997.
[39] Skowron.A.The discernibility matrices and functinns in Information system. R.. Slowinski. Intelligent decision support.Handbook of Applications and Advances of the rough set theory[J]. Kluwer Academic Publishers.1992,331~363
[40] 张文修等,编著.粗糙集理论与方法[M].北京:科学出版社,2001
[41] Tengfei Zhang,Jianmei Xiao.Dynamic System Modeling based on Rough Sets and RBF Neural Networks.Proceedings of the 5th World Congress on Intelligent Control and Automation[J], June 15-19,2004,Hangzhou,P.R.China,pp.185-189
[42] 王东锴,梁樑．含冲突的决策表中的决策规则发现[J].系统工程,2003.vol.21,No.l
[43] V.N.Vapnik.The Nature of Statistical Learning Theory[J].SPringer New York,1995
[44] 张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(l):32 一 42
[45] C.Burges.A tutorial on support vector machines for pattern recognition.Data Mining and Knowledge Discovery,(2)1998:121~ 167
[46] 边肇祺,张学工.模式识别[M],清华大学出版社,北京,2000
[47] 余建平.基于改进神经网络的污水处理参数软测量模型研究[D].重庆:重庆大学城市建设与环境工程学院,2005
[48] 卿晓霞,余建平,王波.基于粗糙集-神经网络的污水参数软测量技术研究[J],2005.3.
[49] 王国胤.Rough 集理论与知识获取[M],西安交通大学出版社.
[50] 张自杰,林荣忱,金儒霖.排水工程[M].第四版.北京:中国建筑工业出版社,2006.6,103～107.
[51] 奚旦立,孙裕生,刘秀英.环境检测[M].第二版.北京:高等教育出版社,2000,427～465
[52] Chih-Wei Hsu, Chih-Chung Chang, and Chih-Jen Lin. A Practical Guide to Support Vector Classification.http://www.csite.ntu.edu.tw.
[53] Hsuan-Tien Lin and Chih-Jen Lin. A Study on Sigmoid Kernels for SVM and Training of non-PSD Kernels by SMO-type Methods.http://www.csie.ntu.edu.tw/~cjlin/papers/tanh.pdf
[54] 李海青,黄志尧等.软测量技术原理及应用[M].化学工业出版社.2000.9
[55] 张文艺,钟梅英等.活性污泥法系统人工神经网络模型.给水排水,2002,28(6):12~15
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |