基于改进极限学习机的光谱定量建模方法

周美灵; 郑德智; 娄格; 刘峥

doi:10.13700/j.bh.1001-5965.2016.0459

基于改进极限学习机的光谱定量建模方法

doi: 10.13700/j.bh.1001-5965.2016.0459

北京航空航天大学仪器科学与光电工程学院, 北京 100083

基金项目:

国家科技支撑计划 2014BAF08B01

详细信息

作者简介:
周美灵, 女, 硕士研究生。主要研究方向:近红外光谱分析

郑德智, 男, 博士, 副教授, 博士生导师。主要研究方向:传感器敏感机理及检测系统

通讯作者:
郑德智, E-mail:zhengdezhi@buaa.edu.cn

中图分类号: O433.4;TB96
计量
- 文章访问数: 1016
- HTML全文浏览量: 205
- PDF下载量: 512
- 被引次数: 0
出版历程
- 收稿日期: 2016-05-31
- 录用日期: 2016-07-01
- 网络出版日期: 2017-06-20

Quantitative spectral modeling method based on improved extreme learning machine

School of Instrumentation Science and Opto-electronics Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100083, China

Funds:

National Key Technology Research and Development Program of China 2014BAF08B01

More Information

Corresponding author: ZHENG Dezhi, E-mail:zhengdezhi@buaa.edu.cn

摘要

摘要:
依据近红外光谱（NIR）产生原理，提出了粒子群优化极限学习机（PSO-ELM）算法，运用于小样本氨水浓度定量分析。通过优化极限学习机（ELM）隐藏节点参数，解决了极限学习机由于输入权值和隐含层偏差随机产生的建模结果具有随机性的问题，提高了预测模型的稳定性、精确度和泛化性能。经实验验证，优化后的PSO-ELM相比ELM，模型预测集均方根误差由0.011 66减小至0.003 22，预测集相关系数由0.995 1提高至0.997 9。将优化后的模型预测结果与支持向量机（SVM）、BP神经网络算法等传统方法的建模结果进行对比，优化后的PSO-ELM算法具有较高的精确度和良好的泛化性能，模型预测效果优于传统的定量回归分析算法。
- 近红外光谱(NIR) /
- 定量分析 /
- 小样本 /
- 极限学习机(ELM) /
- 粒子群优化(PSO)算法
Abstract:
According to the principle of near infrared spectrum (NIR), an optimized extreme learning machine algorithm with particle swarm optimization (PSO-ELM) was proposed and used in ammonia concentration quantitative analysis of small sample. By optimizing the hidden node parameters of extreme learning machine (ELM) algorithm, the problem of randomly generated input weight and hidden layer of ELM leading to random modeling results has been solved. At the same time, the model stability, accuracy and generalization performance were improved. Through the experimental verification, by the optimized PSO-ELM, compared to ELM, the root mean square error of prediction set reduces to 0.003 22 from 0.011 66 and the correlation coefficient of prediction increases from 0.995 1 to 0.997 9. After comparing the optimized model prediction results with the modeling results of traditional support vector machine (SVM) regression and BP neural network algorithm, optimized PSO-ELM offers high accuracy and excellent generalization performance. Model prediction effect is superior to the traditional quantitative regression analysis algorithm.
- near infrared spectrum (NIR) /
- quantitative analysis /
- small sample /
- extreme learning machine (ELM) /
- particle swarm optimization (PSO) algorithm

HTML全文

图 1 氨分子振动模式

Figure 1. Vibration modes of ammonia molecule

下载: 全尺寸图片幻灯片

图 2 氨水近红外吸收光谱

Figure 2. Near infrared absorption spectrum of ammonia

下载: 全尺寸图片幻灯片

图 3 典型单隐含层前馈神经网络结构

Figure 3. A typical single-hidden layer feedforward neural network structure

下载: 全尺寸图片幻灯片

图 4 PSO-ELM算法流程图

Figure 4. Flowchart of PSO-ELM algorithm

下载: 全尺寸图片幻灯片

图 5 氨水近红外光谱预处理前后对比

Figure 5. Comparison of ammonia near infrared spectrum before and after preprocessing

下载: 全尺寸图片幻灯片

图 6 4种算法训练集拟合结果

Figure 6. Imitative effects of training sets of four algorithms

下载: 全尺寸图片幻灯片

图 7 4种算法预测集拟合结果

Figure 7. Imitative effects of prediction sets of four algorithms

下载: 全尺寸图片幻灯片

表 1 训练集拟合效果对比

Table 1. Comparison of imitative effects of training sets

算法类型	RMSEC	R_c
BP-ANN/SVM	0.032 783	0.997 3
ELM/PSO-ELM	0.001 543	0.999 6

下载: 导出CSV

表 2 4种算法建模结果对比

Table 2. Comparison of modeling results among four algorithms

算法类型	氨水浓度		浓度误差	RMSEP	R_p
算法类型	真实值	预测值	浓度误差	RMSEP	R_p
BP-ANN	0.04	0.064 23	0.024 23	0.014 28	0.993 8
	0.09	0.092 75	0.002 75
	0.14	0.145 53	0.005 33
	0.19	0.179 77	-0.010 23
	0.24	0.220 36	-0.019 64
SVM	0.04	0.040 84	0.000 84	0.014 55	0.980 7
	0.09	0.092 72	0.002 72
	0.14	0.130 62	-0.009 38
	0.19	0.189 76	-0.000 24
	0.24	0.207 74	-0.032 26
ELM	0.04	0.013 72	-0.026 28	0.011 66	0.995 1
	0.09	0.103 68	0.013 68
	0.14	0.136 14	-0.003 86
	0.19	0.189 76	-0.000 24
	0.24	0.231 55	-0.008 45
PSO-ELM	0.04	0.042 64	0.002 64	0.003 22	0.997 9
	0.09	0.085 39	-0.004 61
	0.14	0.138 42	-0.001 58
	0.19	0.194 34	0.004 34
	0.24	0.238 40	-0.001 60

下载: 导出CSV

参考文献(23)

[1]	王海莲, 万向元, 胡培松, 等.稻米脂肪含量近红外光谱分析技术研究[J].中国农业科学, 2005, 38(8):1540-1546. http://www.cnki.com.cn/Article/CJFDTOTAL-ZNYK200508005.htm WANG H L, WAN X Y, HU P S, et al.Quantitative analysis of fat content in brown rice by near infrared spectroscopy (NIRS) technique[J].Scientia Agricultura Sinica, 2005, 38(8):1540-1546(in Chinese). http://www.cnki.com.cn/Article/CJFDTOTAL-ZNYK200508005.htm
[2]	吴琼, 原忠虎, 王晓宁.基于偏最小二乘回归分析综述[J].沈阳大学学报, 2007, 19(2):33-35. http://www.cnki.com.cn/Article/CJFDTOTAL-SYDA200702007.htm WU Q, YUAN Z H, WANG X N.Summary of partial least squares regression[J].Journal of Shenyang University, 2007, 19(2):33-35(in Chinese). http://www.cnki.com.cn/Article/CJFDTOTAL-SYDA200702007.htm
[3]	CORTES C, VAPNIK V.Support-vector networks[J].Machine Learning, 1995, 20(3):273-297.
[4]	SUYKENS J A K, VANDEWALLE J.Least squares support vector machine classifiers[J].Neural Processing Letters, 1999, 9(3):293-300. doi: 10.1023/A:1018628609742
[5]	HEERMANN P D, KHAZENIE N.Classification of multispectral remote sensing data using a back-propagation neural network[J].IEEE Transactions on Geoscience and Remote Sensing, 1992, 30(1):81-88. doi: 10.1109/36.124218
[6]	HUANG G B, ZHU Q Y, SIEW C K.Extreme learning machine:A new learning scheme of feedforward neural networks[C]//2004 IEEE International Joint Conference on Neural Networks, 2004.Piscataway, NJ:IEEE Press, 2004, 2:985-990.
[7]	MASRI D, WOON W L, AUNG Z.Soil property prediction:An extreme learning machine approach[C]//22nd International Conference on Neural Information Processing, ICONIP 2015.Berlin:Springer, 2015:18-27.
[8]	GUO W, GU J, LIU D, et al.Peach variety identification using near-infrared diffuse reflectance spectroscopy[J].Computers and Electronics in Agriculture, 2016, 123(C):297-303. https://www.researchgate.net/publication/298732923_Peach_variety_identification_using_near-infrared_diffuse_reflectance_spectroscopy
[9]	HUANG G B, CHEN L.Enhanced random search based incremental extreme learning machine[J].Neurocomputing, 2008, 71(16):3460-3468. https://www.researchgate.net/publication/222679318_Enhanced_random_search_based_incremental_extreme_learning_machine
[10]	JIN Y, LI J, LANG C, et al.Multi-task clustering ELM for VIS-NIR cross-modal feature learning[J].Multidimensional Systems and Signal Processing, 2016, 28(3):1-16. https://www.researchgate.net/publication/301335230_Multi-task_clustering_ELM_for_VIS-NIR_cross-modal_feature_learning
[11]	ALEXANDRE E, CUADRA L, SALCEDO-SANZ S, et al.Hybridizing extreme learning machines and genetic algorithms to select acoustic features in vehicle classification applications[J].Neurocomputing, 2015, 152:58-68. doi: 10.1016/j.neucom.2014.11.019
[12]	陈媛媛, 张记龙, 赵冬娥.基于极限学习机的混合气体FTIR光谱定量分析[J].中北大学学报(自然科学版), 2011, 32(5):636-641. http://www.cnki.com.cn/Article/CJFDTOTAL-HBGG201105023.htm CHEN Y Y, ZHANG J L, ZHAO D E.Quantitative analysis of mixed gas FTIR spectrum based on extreme learning machine[J].Journal of North University of China (Natural Science Edition), 2011, 32(5):636-641(in Chinese). http://www.cnki.com.cn/Article/CJFDTOTAL-HBGG201105023.htm
[13]	DE OLIVEIRA J F L, LUDERMIR T B.Homogeneous ensemble selection through hierarchical clustering with a modified artificial fish swarm algorithm[C]//2011 23rd IEEE International Conference on Tools with Artificial Intelligence (ICTAI).Piscataway, NJ:IEEE Press, 2011:177-180.
[14]	WORKMAN J, WEYER L. 近红外光谱解析实用指南[M]. 褚小立, 许育鹏, 田高友, 译. 北京: 化学工业出版社, 2009: 6-17. WORKMAN J, WEYER L. Practical guide to interpretive near-infrared spectroscopy[M].CHU X L, XU Y P, TIAN G Y, translated.Beijing:Chemical Industry Press, 2009:6-17(in Chinese).
[15]	严衍禄.近红外光谱分析基础与应用[M].北京:中国轻工业出版社, 2005:128-130. YAN Y L.Near infrared spectrum analysis and application[M].Beijing:China Light Industry Press, 2005:128-130(in Chinese).
[16]	POGGIO T, GIROSI F.A theory of networks for approximation and learning[M].Cambridge:Massachusetts Institute of Technology, 1989:25-39.
[17]	陈林伟. 基于极限学习机的骨髓细胞识别技术研究[D]. 杭州: 中国计量学院, 2014: 44-49. CHEN L W.Research on recognition of bone marrow cells based on extreme learning machine[D].Hangzhou:China Jiliang University, 2014:44-49(in Chinese).
[18]	KENNEDY J, EBERHART R.Particle swarm optimization[J].Swarm Intelligence, 2007, 1(1):33-57. doi: 10.1007/s11721-007-0002-0
[19]	石山, 刘德鹏, 李成茂.基于改进粒子群优化算法的飞机作动系统功率调度[J].北京航空航天大学学报, 2016, 42(10):2024-2030. http://bhxb.buaa.edu.cn/CN/abstract/abstract13798.shtml SHI S, LIU D P, LI C M.Power dispatch of actuator of aircraft based on improved particle swarm optimization algorithm[J].Journal of Beijing University of Aeronautics and Astronautics, 2016, 42(10):2024-2030(in Chinese). http://bhxb.buaa.edu.cn/CN/abstract/abstract13798.shtml
[20]	李永强. 基于粒子群优化的极限学习机的XML文档分类中的研究与应用[D]. 沈阳: 东北大学, 2013: 27-33. LI Y Q.Research and application of XML classification based on extreme learning machine with particle swarm optimization[D].Shenyang:Northeastern University, 2013:27-33(in Chinese).
[21]	PRUETT E.Techniques and applications of programmable spectral pattern coding in Texas Instruments DLP spectroscopy[C]//Proceedings of SPIE-The International Society for Optical Engineering.Bellingham, WA:SPIE, 2015.
[22]	刘小丽. 红外光谱在药物分析中的应用研究[D]. 西安: 西北大学, 2013: 6-8. LIU X L.The application of infrared spectroscopy in pharma-ceutical analysis[D].Xi'an:Northwest University, 2013:6-8(in Chinese).
[23]	GORRY P A.General least-squares smoothing and differentiation by the convolution (Savitzky-Golay) method[J].Analytical Chemistry, 1990, 62(6):570-573. doi: 10.1021/ac00205a007