Terahertz Spectrum Features Extraction Based on Kernel Optimization Relevance Vector Machine
ZHONG Yi-wei1, SHEN Tao1,2*, MAO Cun-li1, YU Zheng-tao1
1. School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, China 2. School of Materials Science and Engineering, Kunming University of Science and Technology, Kunming 650500, China
Abstract：Terahertz spectrum is sensitive to the change of the nonlocal molecular vibration mode. Accordingly, the spectral waveform is susceptible to variety of physical and chemical factors, which will lead to peak changes, frequency shifts, and even deformation of the overall waveform. Component analysis and material identification from the correspondence between the fixed peak features and materials will prone to cause errors or mistakes. Therefore, to solve this problem, we proposed a method based on Kernel Optimization Relevance Vector Machine (KO-RVM), which extracts global graphic features to distinct from the local features extraction method. And we use Support Vector Regression (SVR) algorithm as comparison. The result shows that, when basis functions’ parameters of RVM are optimized with expectation-maximization algorithm, it will be suitable for feature extraction of terahertz transmission spectrum. The spectrum can be sparsely represented, and the amount of extracted graphic features is substantially reduced. Reconstruction models based on these features are capable of retaining the overall spectral characteristics, and fitting results for each band are more consistent, while the extracted spectrum features can be used as basis of similarity measurement and the common characteristics investigation between different materials.
Key words：Terahertz frequency spectrum;Feature extraction;Relevance vector machine;Kernel optimize
[1] Baxter J, Guglietta G. Analytical Chemistry, 2011, 83：4342. [2] Fuse N, Takahashi T，Ohki Y, et al. IEEE Electrical Insulation Magazine, 2011, 27(3): 26. [3] Ji T, Zhao H, Han P, et al. Nuclear Science and Techniques, 2013, 24(1): 1. [4] Burnett A, Kendrick J, Russell C, et al. Analytical Chemistry, 2013, 85(16): 7926. [5] Li X, Fu X, Liu J, et al. Journal of Molecular Structure, 2013, 1049: 441. [6] King M，Buchanan W，Korter T. Analytical Chemistry, 2011, 83(10): 3786. [7] Ueno Y, Rungsawang R, Tomita I, et al. Analytical Chemistry, 2006, 78(15): 5424. [8] Kim J, Boenawan R, Ueno Y, et al. Journal of Lightwave Technology, 2014, 32(20): 3768. [9] Ermolina I, Darkwah J, Smith G. AAPS Pharmscitech, 2014, 15(2): 253. [10] Dutta P, Tominaga K. Journal of Molecular Liquids, 2009, 147(1-2): 45. [11] Chen T, Li Zhi, Mo Wei. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2013, 106: 48. [12] Ge H, Jiang Y, Xu Z, et al. Optics Express, 2014, 22(10): 12533. [13] Avila F, Mora M, Oyarce M, et al. Journal of Food Engineering, 2015, 162: 9. [14] El Haddad J, de Miollis F, Sleiman J, et al. Analytical Chemistry, 2014, 86(10): 4927. [15] Wang Q, Ma Y. Chemometrics and Intelligent Laboratory Systems, 2013, 127: 43. [16] Tipping M. Journal of Machine Learning Research, 2001，(1): 211. [17] Vapnik V. IEEE Transactions on Neural Networks, 1999, 10(5): 988. [18] Bowd C, Medeiros F, Zhang Z, et al. Investigative Ophthalmology & Visual Science, 2005, 46(4): 1322. [19] Tipping M, Faul A. Fast Marginal Likelihood Maximization for Sparse Bayesian Models. Proceedings of the Ninth International Workshop on Artificial Intelligence and Statistics, 2003. [20] Schmolck A，Everson R. Machine Learning, 2007, 68(2): 107. [21] Mohsenzadeh Y，Sheikhzadeh H. IEEE Transactions on Neural Networks and Learning Systems, 2015, 26(4): 709. [22] Cheng D, Nguyen M, Gao J, et al. Neural Networks, 2013, 48: 173.
