Supervised Manifold Learning Method Based on Local Feature Extraction

doi:10.3901/JME.2017.24.211

Abstract

Abstract: A supervised manifold learning method based on local feature extraction is proposed to improve the quality of industrial product. A new neighborhood search based on multi-class is used to obtain neighborhood matrices. The eigenvalue decomposition is applied to extract the manifold hidden in the data from the neighborhood matrices. At the same time, a monitoring model is built with the training data based on support vector data description (SVDD). If an abnormal sample is detected by SVDD, it will be projected on the manifold to obtain the adjustment values, which make the abnormal sample return to the normal state. The actual production process data of IF steel is conducted to verify the effectiveness of the proposed method. The results show that the new method can extract the essential manifold and optimize the process parameters effectively. So the proposed method can provide a new strategy to optimize the process parameters for the actual production process.

Key words: local feature extraction, parameters optimization, product quality modeling, supervised manifold learning

CLC Number:

TP277

LI Min, YANG Mengyao, CHEN Ze, ZHAO Qidong. Supervised Manifold Learning Method Based on Local Feature Extraction[J]. Journal of Mechanical Engineering, 2017, 53(24): 211-220.

References

[1] 张绍辉, 李巍华. 基于特征空间降噪的局部保持投影算法及其在轴承故障分类中的应用[J]. 机械工程学报, 2014, 50(3):92-99. ZHANG Shaohui, LI Weihua. Locality preserving projections based on feature space denoising and its application in bearing fault classification[J]. Chinese Journal of Mechanical Engineering, 2014, 50(3):92-99.
[2] TENENBAUM J, SILVA V, LANGFORD J. A global geometric framework for nonlinear dimensionality reduction[J]. Science, 2000, 290(5500):2319-2323.
[3] ROWEIS S, SAUL L. Nonlinear dimensionality reduction by locally linear embedding[J]. Science, 2000, 290(5500):2323-2326.
[4] BELKIN M, NIYOGI P. Laplacian eigenmaps for dimensionality reduction and data representation[J]. Neural Computation, 2003, 15(15):1373-1396.
[5] ZHANG Zhenyue, ZHA Hongyuan. Principal manifolds and nonlinear dimensionality reduction via local tangent space alignment[J]. SIAM Journal of Scientific Computing, 2004, 1(26):313-338.
[6] HE Xiaofei, YAN Shuicheng, HU Yuxiao, et al. Face recognition using laplacian faces[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 3(27):328-340.
[7] ZHANG Tianhao, YANG Jie, ZHAO Deli, et al. Linear local tangent space alignment and application to face recognition[J]. Neurocomputing, 2007, 70(7-9):1547-1553.
[8] LAW M, JAIN A. Incremental nonlinear dimensionality reductionby manifold learning[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 3(28):337-391.
[9] GAO Quanxue, XU Hui, LI Yiying, et al. Two-dimensional supervised local similarity and diversity projection[J]. Pattern Recognition, 2010, 10(43):3359-3363.
[10] YANG Wankou, SUN Changyin, ZHANG Lei. A multi-manifold discriminant analysis method for image feature extraction[J]. Pattern Recognition, 2011, 8(44):1649-1657.
[11] BOZAS K, IZQUIERDO E. Discriminant pairwise local embeddings[C]//IEEE International Conference on Multimedia & Expo Workshops, 2013, 370:1-4.
[12] ORSENIGO C, VERCELLIS C. Kernel ridge regression for out-of-sample mapping in supervised manifold learning[J]. Expert Systems with Applications, 2012, 9(39):7757-7762.
[13] RADUCANU B, DORMAIKA F. A supervised non-linear dimensionality reduction approach for manifold learning[J]. Pattern Recognition, 2012, 45(6):2432-2444.
[14] HUA Qiang, BAI Lijie, WANG Xizhao, et al. Local similarity and diversity preserving discriminant projection for face and handwriting digits recognition[J]. Neurocomputing, 2012, 86(4):150-157.
[15] GAO Quanxue, MA Jingjie, ZHANG Hailin, et al. Stable orthogonal local discriminant embedding for linear dimensionality reduction[J]. IEEE Transactions on Image Processing, 2013, 22(7):2521-2531.
[16] RADUCANU B, DORMAIKA F. Embedding new observations via sparse-coding for non-linear manifold learning[J]. Pattern Recognition, 2014, 47(1):480-492.
[17] 王萌,孙树栋.基于优化核空间的制造过程质量分析算法[J]. 机械工程学报,2012, 48(22):182-188. WANG Meng, SUN Shudong. Optimized kernel space based algorithm for quality data analysis[J]. Journal of Mechanical Engineering, 2012, 48(22):182-188.
[18] YU Jianbo. Semiconductor manufacturing process monitoring using gaussian mixture model and Bayesian method with local and nonlocal information[J]. IEEE Transactions on Semiconductor Manufacturing, 2012, 25(25):480-493.
[19] LIU Yuanjui, CHEN Tao, YAO Yuan. Nonlinear process monitoring and fault isolation using extended maximum variance unfolding[J]. Journal of Process Control, 2014, 24(6):880-891.
[20] XIAO Zhibo, WANG Huangang, ZHOU Junwu. Robust dynamic process monitoring based on sparse representation preserving embedding[J]. Journal of Process Control, 2016, 40:119-133.
[21] TONG Chudong, SHI Xuhua, LAN Ting. Statistical process monitoring based on orthogonal multi-manifold projections and a novel variable contribution analysis[J]. ISA Transactions, 2016, 65:407.
[22] JOHN S. Kernel methods for pattern analysis[M]. Cambridge:Cambridge University Press, 2006.
[23] 中华人民共和国国家质量监督检验检疫总局,中国国家标准化管理委员会.GB/T 5213-2008冷轧低碳钢板及钢带[S].北京:中国标准出版社,2009. General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China, Standardization Administration of the Republic of China. GB/T 5213-2008 Cold rolled low carbon steel sheet and strip[S]. Beijing:Standards Press of China, 2009.