Novel Variable Selection Method Based on Uninformative Variable Elimination and Ridge Extreme Learning Machine: CO Gas Concentration Retrieval Trial
CHEN Yuan-yuan1,2,3, WANG Zhi-bin1,2,3, WANG Zhao-ba1,2,3
1. State Key Laboratory for Electronic Measurement Technology, North University of China, Taiyuan 030051, China 2. Key Lab of Instrumentation Science & Dynamic Measurement, North University of China, Ministry of Education, Taiyuan 030051, China 3. Engineering Technology Research Center of Shanxi Province for Opto-Electronic Information and Instrument, North University of China, Taiyuan 030051, China
Abstract:Variable selection is an essential part in spectroscopy analysis area. To overcome the problems of traditional interval selection methods, this paper proposed a novel variable selection and assessment method based on uninformative variable elimination (UVE) and ridge extreme learning machine (RELM) algorithms. Firstly, the UVE method was adopted to eliminate the uninformative wavelengths. Secondly, to solve the collinearity problem, RELM algorithm was adopted to replace the traditional modeling methods (PLS, BP neural network, etc.). Finally, the optimal combination of wavelength regions was selected by using feature selection path (FSP) plot and sparsity-error trade-off (SET) curve. The experiment results of CO gas concentration retrieval showed that (1) the UVE algorithm can select the most informative variables, which were the feature wavelengths of the CO gas transmittance spectrum; (2) the RELM algorithm has the advantage of rapid modeling, solving the collinearity problem, and high accuracy (the determined coefficient r of CO gas concentration retrieval can reach 0.995); (3) the FSP plot and SET curve were easy understanding, also intuitive to experts to find the best combination of wavelengths and extract useful domain knowledge.
[1] Z. Xiaobo, Z. Jiewen, Povey M J, et al. Anal. Chim. Acta, 2010, 667(1-2): 14. [2] Javier Moros J K, Guillermo Quintas, Salvador Garrigues,et al. Anal. Chim. Acta, 2008, 630(2): 150. [3] Frénay Benot, Mark van Heeswijk, Yoan Miche,et al. Neurocomputing, 2013, 102: 111. [4] Huang G B, Zhu Q Y, Siew C K. Neurocomputing, 2006, 70(1-3): 489. [5] Li G Q, Niu P F. Neural Comput. Appl., 2013, 22: 803. [6] Araújo M C U, Saldanha T C B, Galvao R K H, et al. Chemometr. Intell. Lab, 2001, 57(2): 65. [7] Liu F, Jiang Y, He Y. Anal. Chim. Acta, 2009, 635(1): 45. [8] Pontes M J C, Cortez J, Galvao R K H, et al. Analytica Chimica Acta, 2009, 642(1-2): 12. [9] Ouyang A G, Liu J. Meas. Sci. Technol., 2013, 24. [10] Centner V, Massart D L, deNoord O E, et al. Anal. Chem., 1996, 68(21): 3851. [11] Abrahamsson C, Johansson J, Sparen A, et al. Chemometr. Intell. Lab, 2003, 69(1-2): 3. [12] Cai W S, Li Y K, Shao X G. Chemometr. Intell. Lab, 2008, 90(2): 188. [13] Todeschini R, Galvagni D, Vilchez J L,et al. Trac-Trend. Anal. Chem., 1999, 18(2): 93. [14] Boger Z. Analytica Chimica Acta, 2003, 490: 31. [15] Lucasius C B, Beckers M L M, Kateman G. Analytica Chimica Acta, 1994, 286(2): 135. [16] Leardi R, Gonzalez A L. Chemometr. Intell. Lab, 1998, 41: 195. [17] Leardi R, Seasholtz M B, Pell R J. Analytica Chimica Acta, 2002, 461(2): 189. [18] Ghasemi J, Niazi A, Leardi R. Talanta, 2003, 59(2): 311. [19] Abdollahi H, Bagheri L. Analytica Chimica Acta, 2004, 514(2): 211. [20] Gourvenec S, Capron X, Massart D L. Analytica Chimica Acta, 2004, 519(1): 11. [21] Durand A, Devos O, Ruckebusch C. Analytica Chimica Acta, 2007, 595(1-2): 72. [22] da Costa Filho P A. Anal. Chim. Acta, 2009, 631(2): 206. [23] Givianrad M H, Saber-Tehrani M, Zarin S,et al. Journal of the Serbian Chemical Society, 2013, 78(4): 555. [24] Norgaard L, Saudland A, Wagner J, et al. Appl. Spectrosc., 2000, 54(3): 413. [25] Pereira A F C, Pontes M J C, Gambarra F F, et al. Food Res. Int., 2008,41: 341. [26] Cramer J A, Kramer K E, Johnson K J, et al. Chemometr. Intell. Lab, 2008, 92(1): 13. [27] Zou X B, Zhao J W, Huang X Y, et al. Chemometr. Intell. Lab, 2007, 87(1): 43. [28] Zou X B, Zhao H W, Li Y X. Vib. Spectrosc., 2007, 44: 220. [29] Hemmateenejad B, Akhond M, Samari F. Spectrochim Acta A, 2007, 67(3-4): 958. [30] Chen D, Cai W S, Shao X G. Chemometr. Intell. Lab, 2007, 87: 312. [31] Bogomolov A, Hachey M. Chemometr. Intell. Lab, 2007, 88(1): 132. [32] Kasemsumran S, Du Y P, Maruo K, et al. Chemometr. Intell. Lab, 2006,82:97. [33] Norgaard L, Hahn M T, Knudsen L B, et al. Int. Dairy J, 2005, 15(12): 1261. [34] Stordrange L, Rajalahti T, Libnau F O. Chemometr. Intell. Lab, 2004, 70(2): 137. [35] Du Y P, Liang Y Z, Jiang J H, et al. Analytica Chimica Acta, 2004, 501(2): 183. [36] Baskir I M, Drozd A V. Chemometr. Intell. Lab, 2003, 66(1): 89. [37] Jiang J H, Berry R J, Siesler H W, et al. Anal. Chem., 2002, 74(14): 3555.