|
|
|
|
|
|
Research on Classification of Dwarf Nova Based on Deep Architecture Network |
ZHAO Yong-jian, GUO Rui, WANG Lu-yao, JIANG Bin* |
School of Mechanical, Electrical & Information Engineering, Shandong University, Weihai, Weihai 264209, China |
|
|
Abstract Dwarf nova (DN) is a special and rare class of semi-contiguous binary star. To discovery more DNs is significant for the further study of matter transfer theory. It also has been profound for understanding the evolution of close binary stars. It is a research hot spot to extract features of celestial spectra and then classify them by deep learning. Traditional auto-encoder is a classical neural network model with only one hidden layer. However, its coding ability is limited and data representation learning ability is insufficient. Broadening the depth of the neural network with modularity can make the network learn features of the celestial spectrum successively. High-level features can be obtained through gradual abstract learning of underlying features so as to improve the spectral classification accuracy. In this paper, a deep feedforward stack network is constructed consisting of an input layer, several hidden layers and an output layer on the basis of auto-encoder. This network with multi-layer perceptron architecture is utilized to process massive spectral data sets. It excavates the depth structure features hidden in the spectra and realizes the accurate classification of DN spectra. Parameters set for the network with deep architecture will seriously affect the performance of the constructed network. In this paper, the optimization of network parameters is divided into two processes: hierarchical training and inverse propagation. The preprocessed spectral data first enter the network from the input layer, and then the network parameters are trained layer by layer with the auto-encoder algorithm and weight sharing policy. In the reverse propagation stage, the initial sample data are input into the network again, and the network is initialized with the weights obtained from the hierarchical training process. Then the local optimization training results of each layer are fused and the network parameters are adjusted according to the set output error cost function. Hierarchical training and inverse propagation are operated repeatedly until the global optimal network parameters are obtained. Finally, the last hidden layer is adopted as the reconstruction layer to connect the support vector machine classifier, and the feature extraction and classification of DNs are realized. In the process of network parameter optimization, the idea of mean network is utilized to make the output of network hidden layer unit attenuate according to dropout coefficient. The reverse propagation algorithm is adopted to fine-tune the entire network to prevent depth overfitting in the network. Such operation can reduce to extract duplicate feature caused by mutual moderation of hidden layer neurons, and improve the generalization ability. The distributed multi-layer architecture of the network can provide effective data abstraction and representational learning. The feature detection layer can learn the depth structure features implicitly from the unlabeled data, and effectively characterize the nonlinearity and random fluctuation of spectral data, thus avoiding the explicit extraction of spectral features. The network shows strong data fitting and generalization ability. Weight sharing between different layers can reduce the interference of redundant information and effectively resolve the risk that the traditional multi-layer architecture network is prone to fall into the local minimization of weight. Experiments show that that of the accuracy of the deep architecture network in DNs classification is 95.81%, higher than the classical LM-BP network.
|
Received: 2019-01-09
Accepted: 2019-04-12
|
|
Corresponding Authors:
JIANG Bin
E-mail: jiangbin@sdu.edu.cn
|
|
[1] Zhao G, Zhao Y H, et al. Research in Astronomy and Astrophysics, 2012, 12: 723.
[2] Bolton A S, Schlegel D J, et al. The Astronomical Journal, 2012, 144(5): 144.
[3] Wei P, Luo A L, Li Y B. The Astronomical Journal, 2014, 147(5): 101.
[4] Rebassa M A, et al. Monthly Notices of the Royal Astronomical Society, 2011, 402: 620.
[5] Daniel S F, Connolly A, Schneider J. The Astronomical Journal, 2011, 142(6): 203.
[6] Bengio Y. Neural Network: Tricks of the Trade, 2012, 7700: 437.
[7] WANG Wen-yu, GUO Ge-lin, JIANG Bin, et al(王文玉,郭格霖,姜 斌,等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2016, 36(10): 3360.
[8] Bazarghan M. Astrophysics and Space Science, 2012, 337(1): 93.
[9] Krizhevsky A, Sutskever I. Advances in Neural Information Processing Systems, 2012, 25(2): 2012.
[10] Oommen B J, Thomas A. Pattern Recognition, 2014, 47(1): 40. |
[1] |
YANG Qun1, 2, LING Qi-han1, WEI Yong1, NING Qiang1, 2, KONG Fa-ming1, ZHOU Yi-fan1, 2, ZHANG Hai-lin1, WANG Jie1, 2*. Non-Destructive Monitoring Model of Functional Nitrogen Content in
Citrus Leaves Based on Visible-Near Infrared Spectroscopy[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(11): 3396-3403. |
[2] |
HUANG Zhao-di1, CHEN Zai-liang2, WANG Chen3, TIAN Peng2, ZHANG Hai-liang2, XIE Chao-yong2*, LIU Xue-mei4*. Comparing Different Multivariate Calibration Methods Analyses for Measurement of Soil Properties Using Visible and Short Wave-Near
Infrared Spectroscopy Combined With Machine Learning Algorithms[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(11): 3535-3540. |
[3] |
LIU Shu1, JIN Yue1, 2, SU Piao1, 2, MIN Hong1, AN Ya-rui2, WU Xiao-hong1*. Determination of Calcium, Magnesium, Aluminium and Silicon Content in Iron Ore Using Laser-Induced Breakdown Spectroscopy Assisted by Variable Importance-Back Propagation Artificial Neural Networks[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(10): 3132-3142. |
[4] |
ZHAO Zhi-lei1, 2, 3, 4,WANG Xue-mei1, 2, 3,LIU Dong-dong1, 2, 3,WANG Yan-wei1, 2, 3,GU Yu-hong5,TENG Jia-xin1,NIU Xiao-ying1, 2, 3, 4*. Quantitative Analysis of Soluble Solids and Titratable Acidity Content in Angeleno Plum by Near-Infrared Spectroscopy With BP-ANN and PLS[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2022, 42(09): 2836-2842. |
[5] |
CHEN Yi-fan, LI Yun-jing, PENG Miao-miao, YANG Chun-yong*, HOU Jin, CHEN Shao-ping. Improvements of VIS-NIR Spectroscopy Model in the Prediction of TVB-N Using MIV Wavelength Selection[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2020, 40(05): 1413-1419. |
[6] |
YUAN Ru-jun1, 2, 3, WAN Xiong1, 2*, HE Qiang1, 2, 3, WANG Hong-peng1, 2. Research on Olivine Component Analysis Using LIBS Combined with Back-Propagation Algorithm[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2019, 39(12): 3861-3867. |
[7] |
WANG Xue-ji1, 2, HU Bing-liang1*, YU Tao1, LIU Qing-song1, 2, LI Hong-bo1, 2, FAN Yao1. Analysis of Nitrate in Seawater of Wheat Island Based on LLE-BPNN[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2019, 39(05): 1503-1508. |
[8] |
XIONG Xian-qing1, QIAN Shao-ping2, SHENG Kui-chuan2*, HE Yong2, FANG Lu1, WU Zhi-hui1. Predicting Gross Calorific Value of Agricultural Feedstock Based on Proximate/Ultimate Analysis and Visible-Near Infrared Spectroscopy[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2017, 37(05): 1622-1627. |
[9] |
TAN Hong-ying, ZHENG De-zhong, LI Xue, XU Zheng-xia . The Application of THz Spectroscopy and GA-BP in Methanol Concentration Detection[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2016, 36(11): 3752-3757. |
[10] |
ZHAO Zhi-lei, WANG Yan-wei, GONG Dong-jun, NIU Xiao-ying*, CHENG Wei, GU Yu-hong . Discrimination of Plum Browning with Near Infrared Spectroscopy [J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2016, 36(07): 2089-2093. |
[11] |
ZHANG Chu, LIU Fei, KONG Wen-wen, HE Yong* . Application of Near-Infrared Spectroscopy to Distinguish Brands of Soy Milk Powder and Fake Soy Milk Powder [J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2014, 34(07): 1826-1830. |
[12] |
LI Jin-meng, YE Xu-jun*, WANG Qiao-nan, ZHANG Chu, HE Yong . Development of Prediction Models for Determining N Content in Citrus Leaves Based on Hyperspectral Imaging Technology [J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2014, 34(01): 212-216. |
[13] |
XU Li-peng, GE Liang-quan*, GU Yi, LIU Min, ZHANG Qing-xian, LI Fei, LUO Bin . Research on the Application of Principal Component Analysis and Improved BP Neural Network to the Determination of Fe and Ti Contents in Geological Samples[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2013, 33(05): 1392-1396. |
[14] |
NIU Xiao-ying1, SHAO Li-min2, ZHAO Zhi-lei1, ZHANG Xiao-yu1 . Nondestructive Discrimination of Strawberry Varieties by NIR and BP-ANN[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2012, 32(08): 2095-2099. |
[15] |
PANG Tao-tao,YAO Jian-bin,DU Li-ming*. Artificial Neural Networks for the Identification of Infrared Spectra of Ilex Kudingcha[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2007, 27(07): 1336-1339. |
|
|
|
|