رویکرد شورای انتخاب ویژگی بر اساس خوشه‌بندی سلسله‌مراتبی برای حل مشکل داده‌های زايد در بینی الکترونیکی
محورهای موضوعی : مهندسی برق و کامپیوتر
محمدعلی باقری
1
(دانشگاه تربیت مدرس)
غلامعلی منتظر
2
(دانشگاه تربیت مدرس)
کلید واژه: بینی الکترونیکی خوشه‌بندی سلسله‌مراتبی سیستم دسته‌بند چندگانه شورای انتخاب ویژگی,
چکیده مقاله :
وجود داده‌هاي زايد در پاسخ حس‌گرهای بيني الكترونيكي اثر چشم‌گیری در دسته‌بندی بو دارد. برای بهبود صحت دسته‌بندی، می‌توان از سیستم دسته‌بندی چندگانه بر اساس انتخاب چند زیرمجموعه از ویژگی‌ها (به‌جای استفاده از تمام ابعاد بردار ویژگی) استفاده کرد. در این رویکرد که "شورای انتخاب ویژگی" نامیده می‌شود، فرض بر آن است که مجموعه اولیه ویژگی‌ها دارای داده‌هایی زايد بوده و می‌توان با انتخاب زیرمجموعه‌های ویژگی مختلف و سپس ترکیب دسته‌بندهای ایجادشده با این زیرمجموعه‌ها به نتایج دسته‌بندی بهتری رسید. در این مقاله پس از پیش‌پردازش سیگنال اولیه حس‌گرها و حذف نویز سیگنال با استفاده از تحلیل موجک، سیستم دسته‌بند چندگانه با زیرمجموعه‌های ویژگی مختلف طراحی شده است: ویژگی‌های استخراج‌شده از سیگنال گذرای حسگر با روش خوشه‌بندی سلسله‌مراتبی طبقه‌بندی شده و زیرمجموعه‌های مختلف با انتخاب یک ویژگی از هر خوشه ایجاد شده‌اند. این موضوع موجب بهبود تنوع دسته‌بندهای پایه و افزایش کارایی و سرعت دسته‌بندی می‌شود. روش پیشنهادی ابتدا در چند مجموعه داده تراز از مخزن داده UCI آزمون شده و پس از اثبات توانایی آن، در مجموعه داده بویایی حاصل از رایحه سه نوع شیرین‌بیان به کار برده شده است. نتایج حاصل نشان‌دهنده کارایی روش جدید در شناسایی الگوهای بویایی است.
The redundancy problem of sensor response in electronic noses is still remarkable due to the cross-selectivity of chemical gas sensors which can degrade the classification performance. In such situations, a more efficient multiple classifier system can be obtained in random feature space rather than in the original one. Ensemble Feature Selection (EFS) methods assume that there is redundancy in the overall feature set and better performance can be achieved by choosing different subsets of input features for multiple classifiers. By combining these classifiers the higher recognition rate can be achieved. In this paper, we propose a feature subset selection method based on hierarchical clustering of transient features in order to enhance the classifier diversity and efficiency of learning algorithms. Our algorithm is tested on the UCI benchmark data sets and then used to design an odor recognition system. The experimental results of proposed method based on hierarchical clustering feature subset selection and multiple classifier system demonstrate the more efficient classification performance.
[1] H. T. Nagle, R. Gutierrez - Osuna, and S. S. Schiffman, "The how and why of electronic noses," IEEE Spectr, vol. 35, no. 9, pp. 22-31, Sep. 1998.
[2] K. Arshak, E. Moore, G. M. Lyons, F. Harris, and S. Clifford, "A review of gas sensors employed in electronic nose application," Sens. Rev, vol. 24, no. 2, pp. 181-198, Jan. 2004.
[3] R. Gutierrez - Osuna and H. T. Nagle, "A method for evaluating data preprocessing techniques for odor classification with an array of gas sensors," IEEE Trans. Syst. Man, Cybern. B, Cybern., vol. 29, no. 5, pp. 626-632, Oct. 1999.
[4] R. Gutierrez-Osuna, H. T. Nagle, B. Kermani, and S. S. Schiffman, "signal conditioning and pre-processing, in Handbook of Machine Olfaction: Electronic Nose Technology, T. C. Pearce, S. S. Schiffman, H. T. Nagle, and J. W. Gardner, eds., ed Weinheim, Germany: Wiley - VCH, 2002.
[5] R. Gutierrez - Osuna, "Pattern analysis for machine olfaction: a review," IEEE Sensors J., vol. 2, no. 3, pp. 189-202, Jun. 2002.
[6] E. Phaisangittisagul and H. T. Nagle, "Enhancing multiple classifier system performance for machine olfaction using odor-type signatures," Sensors and Actuators B: Chemical, vol. 125, no. 1, pp. 246-253, 16 Jul. 2007.
[7] E. Phaisangittisagul and H. T. Nagle, "Sensor selection for machine olfaction based on transient feature extraction," IEEE Trans. on Instrumentation and Measurement, vol. 57, no. 2, pp. 369-378, Feb. 2008.
[8] J. W. Gardner, P. Boilot, and E. L. Hines, "Enhancing electronic nose performance by sensor selection using a new integer-based genetic algorithm approach," Sens. Actuators B, Chem., vol. 106, no. 1, pp. 114-121, May 2005.
[9] M. Aleixandre, I. Sayago, M. C. Horrillo, M. J. Fernandez, L. Ares, M. Garcia, C. P. Santos, and J. Gutearres, "Analysis of neural networks and analysis of feature selection with genetic algorithm to discriminate among pollutant gas," Sens. Actuators B, Chem., vol. 103, no. 1-2, pp. 122-128, Sep. 2004.
[10] A. Pardo, S. Marco, A. Ortega, A. Perera, T. Sundic, and J. Samitier, "Proceedings of the Conference International Symposium on Olfaction and Electronic Noses ISOEN," London, 2000, pp. 83-88.
[11] C. Li, P. Heinemann, and R. Sherry, "Neural network and Bayesian network fusion models to fuse electronic nose and surface acoustic wave sensor data for apple defect detection," Sensors and Actuators B, vol. 125, no. 1, pp. 301-310, Jul. 2007.
[12] M. Falasconi, M. Pardo, G. Sberveglieri, I. Ricc`o, and A. Bresciani, "The novel EOS835 electronic nose and data analysis for evaluating coffee ripening," Sensors and Actuators B, vol. 110, no. 1, pp. 73-80, Sep. 2005.
[13] M. Skurichina and R. P. W. Duin, "Bagging, boosting and the random subspace method for linear classifiers," Pattern Analysis & Applications, vol. 5, no. 2, pp. 121-135, Jan. 2002.
[14] T. G. Dietterich, "Machine learning research: four current directions," Artificial Intell. Mag., vol. 18, no. 4, pp. 97-136, Sep. 1997.
[15] L. Breiman, "Bagging predictors," Machine Learning, vol. 24, no. 2, pp. 123-140, May 1996.
[16] Y. Freund and R. Schapire, "Experiments with a new boosting algorithm," in Proc. of the Thirteenth International Conf. on Machine Learning, vol. 2, pp. 148-156, Jun. 1996.
[17] T. K. Ho, "The random subspace method for constructing decision forests," IEEE Trans. on Pattem Analysis and Machine Intelligence, vol. 20, no. 8, pp. 832-844, Jan. 1998.
[18] N. C. Oza and K. Tumer, "Input decimation ensembles: decorrelation through dimensionality reduction," in Proc. 2nd Int. Workshop on Multiple Classifier Systems, Lecture Notes in Computer Science, vol. 2096, J. Kittler and F. Roli, Eds., ed, pp. 238-247, 2001.
[19] M. A. Hall, Correlation - Based Feature Selection for Machine Learning, Ph. D. Dissertation, the University of Waikato, Hamilton, New Zealand, 1999.
[20] M. A. Hall and L. A. Smith, "Feature subset selection: a correlation based filter approach," in Intl. Conf. Neural Inform. Processing Intell. Inform. Syst., vol. 2, pp. 855-858, Sep. 1997.
[21] C. L. Blake and C. J. Merz, UCI Repository of Machine Learning Databases, Department of Information and Computer Sciences, University of California, Irvine, US. Available: http://www.ics.uci.edu/~mlearn/MLRepository.html
[22] T. Skov and R. Bro, Three-Way Electronic Nose Data, 2004. http://www.models.kvl.dk/research/data/3Dnosedata/index.asp
[23] C. S. Burrus, R. A. Gopinath, and H. Guo, Introduction to Wavelets and Wavelet Transforms: A Primer: Prentice Hall: New Jersy, 1998.
[24] S. G. Mallat, "The theory of multiresoloution signal decomposition: the wavelet reperesntation," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 11, no. 7, pp. 674-693, Jan. 1989.
[25] F. Hossein-Babaei and M. Orvatinia, "Gas diagnosis based on selective diffusion retardation in an air filled capillary," Sensors and Actuators B, vol. 96, no. 1-2, pp. 298-303, Dec. 2003.
[26] F. Hossein-Babaei, M. Hemmati, and M. Dehmobed, "Gas diagnosis by a quantitative assessment of the transient response of a capillary - attached gas sensor," Sensors and Actuators B, vol. 107, no. 1, pp. 461-467, May 2005.
[27] S. Balasubramanian, S. Panigrahi, C. M. Logue, H. Gu, and M. Marchello, "Neural networks - integrated metal oxide - based artificial olfactory system for meat spoilage identification," J. of Food Engineering, vol. 91, no. 1-3, pp. 91-98, Apr. 2009.
[28] W. Li, H. Leung, C. Kwan, and B. R. Linnell, "E-nose vapor identification based on Dempster-Shafer fusion of multiple classifiers," IEEE Trans. on Instrumentation and Measurement, vol. 57, no. 10, pp. 2273-2282, Oct. 2008.
[29] B. G. Kermani, On Using Artificial Neural Networks and Genetic Algorithm for Electronic Nose, Ph. D. Dissertation, North Carolina State University, 1996.
[30] B. G. Kermani, S. S. Schiffman, and H. T. Nagle, "Performance of the Levenberg-Marquardt neural network training method in electronic nose applications," Sensors and Actuators B, vol. 110, no. 1, pp. 13-22, Sep. 2005.
[31] J. Fu, G. Li, Y. Qin, and W. J. Freeman, "A pattern recognition method for electronic noses based on an olfactory neural network," Sensors and Actuators B, vol. 125, no. 2, pp. 489-497, Jul. 2007.
[32] Y. Yin, H. Yu, and H. Zhang, "A feature extraction method based on wavelet packet analysis for discrimination of Chinese vinegars using a gas sensors array," Sensors and Actuators B: Chemical, vol. 134, no. 2, pp. 1005-1009, Sep. 2008.
[33] R. Ionescu and E. Llobet, "Wavelet transform - based fast feature extraction from temperature modulated semiconductor gas sensors," Sensors and Actuators B: Chemical, vol. 81, no. 2-3, pp. 289-295, Jan. 2002.
[34] R. Bryll, R. Gutierrez-Osuna, and F. Quek, "Attribute bagging: improving accuracy of classifier ensembles by using random feature subsets," Pattern Recognition, vol. 36, no. 6, pp. 1291-1302, Jun. 2003.
[35] P. Cunningham and J. Carney, "Diversity versus quality in classification ensembles based on feature selectio," in R. L. de Mántaras, E. Plaza (Eds.), Proc. of the ECML 2000, vol. 1810, pp. 109-116, Barcelona, Spain, 2000.