[1]刘凤秋,张红旭.离散神经网络求解支持向量机中的凸优化问题[J].哈尔滨理工大学学报,2018,(04):133-139.[doi:10.15938/j.jhust.2018.04.025]
 LIU Feng qiu,ZHANG Hong xu. A Discretetime Neural Network for Solving Convex Optimization Problem in Support Vector Machine[J].哈尔滨理工大学学报,2018,(04):133-139.[doi:10.15938/j.jhust.2018.04.025]
点击复制

离散神经网络求解支持向量机中的凸优化问题()
分享到:

《哈尔滨理工大学学报》[ISSN:1007-2683/CN:23-1404/N]

卷:
期数:
2018年04期
页码:
133-139
栏目:
管理科学与工程
出版日期:
2018-08-25

文章信息/Info

Title:
 A Discretetime Neural Network for Solving Convex 
Optimization Problem in Support Vector Machine
作者:
刘凤秋张红旭
哈尔滨理工大学 应用科学学院, 黑龙江 哈尔滨 150080
Author(s):
 LIU FengqiuZHANG Hongxu
 School of Applied Sciences, Harbin University of Science and Technology, Harbin 150080, China
关键词:
关键词:离散神经网络 支持向量机 凸优化 全局指数收敛
Keywords:
 Keywords:discretetime neural network support vector machine convex optimization global exponential convergent
分类号:
O23
DOI:
10.15938/j.jhust.2018.04.025
文献标志码:
A
摘要:
摘要:针对正定核函数的支持向量机所导出的凸二次优化问题, 提出了一个离散型神经网络模型。利用KarushKuhnTucker (KKT)条件和投影理论构造投影方程, 使得投影方程的解与优化问题的解一一对应; 进一步基于投影方程建立离散神经网络; 理论结果表明, 网络的平衡点与优化问题的最优解相对应, 且网络具有全局指数收敛性。 相比于连续网络, 所提出的网络结构简单, 减少了计算的复杂度; 所得理论结果保证了网络能够有效求解支持向量机中的优化问题。 最后, 利用分类问题和标准数据集进行实验, 数值结果验证了本文所设计的网络在求解支持向量机优化问题的有效性。
Abstract:
 Abstract:This paper proposes a discretetime neural network model to solve the convex optimization problem deduced by a positivekernelbased support vector machine (SVM)First, the projection equations are constructed through the KarushKuhnTucker (KKT) conditions and projection theory so that there exists a onetoone correspondence between the solution of projection equations and the optimal solution of optimization problem, and then a discretetime neural network was constructed by projection equations Second, the obtained theoretical results indicate that the equilibrium point of the proposed neural network corresponds to the optimal solution of the optimization problem, and the proposed neural network is globally exponentially convergent Compared with some continuous neural networks, the architecture of proposed neural network is simple, which decreases the computational complexity Finally, some classification problems and benchmarking data sets are used in the experiment The numeral results show the efficiency of the proposed neural network for solving the optimization problem in SVM

参考文献/References:

[1]SCHLKOPF B, SMOLA A Learning with kernels[M]. Cambridge : MIT, 2002
[2]CORTES C,VAPNIK V N Support Vector Networks[J] Machine Learning, 1995, 20:273–297
[3]孙永倩, 王培东 基于支持向量机的并行CT图像分割方法[J]. 哈尔滨理工大学学报, 2013, 18(3):42-46
[4]尤波, 周丽娜, 黄玲, 应用于假手的肌电信号分类方法研究[J]. 哈尔滨理工大学学报, 2011, 16(3):1-7
[5]JOACHIMS T Making largescale SVM Learning Practical[C]// Cambridge, Massachusetts, Advances in Kernel Methodssupport Vector Learning, MITPress: Platt J, 1999:169-184
[6]PLATT J Fast Training of Support Vector Machines Using Sequential Minimal Optimization Cambridge[C]// Cambridge, Massachusetts, Advances in Kernel Methodssupport Vector Learning, MITPress: Platt J, 1999: 185–208
[7]MANGASARIAN O, MUSICANT D Lagrangian Support Vector Machines[J]. Journal of Machine Learning Research, 2001(1):161-177
[8]YANG X, SHU L An Extended Lagrange Support Vector Machine for Classification[J]. Progress in Natural Science, 2004,14(6):519-523
[9]ANGUTITA D, BONI A Improved Neural Network for SVM Learning[J]. IEEE Trans on Neural Networks, 2002(13):1243-1244
[10]KG*2〗YANG Y, HE Q A Compact Neural Network for Training Support Vector Machines[J]. Neurocomputing, 2012(86):193-198
[11]KG*2〗ANGUITA D, BONI A Neural Network Learning for Analog VLSI Implementations of Support Vector Machines: A survey[J]. Neurocomputing, 2003(55):265-283
[12]KG*2〗NAZEMI A, DEHGHAN M A Neural Network Method for Solving Support Vector Classification Problems[J]. Neurocomputing, 2015(152):369-376
[13]KG*2〗XIA Y, WANG J A Onelayer Recurrent Neural Network for Support Vector Machine Learning[J]. IEEE Trans on Systems Cybern B, 2004, 34(2):1261-1267
[14]KG*2〗LIU Q, GUO Z, WANG J A Onelayer Recurrent Neural Network for Constrained Pseudoconvex Optimization and its Application for Dynamic Portfolio Optimization[J]. Neural Networks 2012(26):99-109
[15]KG*2〗LIU Q, WANG J A Onelayer Recurrent Neural Network for Constrained Nonsmooth Optimization[J]. IEEE Trans on Systems Cybern B, 2011, 41(5):1323-1332
[16]KG*2〗LIU Q, WANG J A Onelayer Recurrent Neural Network with a Discontinuous Hardlimiting Activation Function for Quadratic Programming[J]. IEEE Trans on Neural Networks, 2008, 19(4):558-568
[17]KG*2〗XU H Projection Neural Networks for Solving Constrained Convex and Degenerate Quadratic Problems[C]// IEEE International Conference on Intelligent Computing & Intelligent Systems, 2010:91-95
[18]KG*2〗XUE X, BIAN W A Project Neural Network for Solving Degenerate Convex Quadratic Program[J]. Neurocomputing, 2007(70):2449-2459
[19]KG*2〗LIU Feng, XUE X Subgradientbased Neural Network for Nonconvex Optimization Problems in Support Vector Machines with Indefinite Kernels[J]. Journal of Industrial & Management Optimization 2016, 12(1):285-301
LL〗[20]KG*2〗PAKE·〗REZLLZARBE M Convergence Analysis of a Discretetime Recurrent Neural Networks to Perform Quadratic Real Optimization with Bound Constraints[J]. IEEE Trans on Neural Networks, 1998, 9(6):1344-1351
[21]KG*2〗TAN K, TANG H, YI Z Global Exponential Stability of Discretetime Neural Networks for Constrained Quadratic Optimization[J]. Neurocomputing, 2004(56):399-406
[22]KG*2〗TANK D, HOPFIELD J Simple Neural Optimization Networks: An A/D Converter, Signal Decision Circuit, and a Linear Programming Circuit[J]. IEEE Trans on Circuit & Systems, 1986(33):533-541
[23]KG*2〗BAZARAA M, SHERALI H, SHETTY C Nonlinear Programming: Theory and Algorithms[M]. New Jersey: John Wiley & Sons, Inc, 1993

相似文献/References:

[1]盖晁旭,梁隆恺,何勇军.数据不充分情况下的说话人识别[J].哈尔滨理工大学学报,2017,(03):13.[doi:10. 15938 /j. jhust. 2017. 03. 003]
 GAI Chao-xu,LIANG Long-kai,HE Yong-jun.Speaker Recognition under Limited Data[J].哈尔滨理工大学学报,2017,(04):13.[doi:10. 15938 /j. jhust. 2017. 03. 003]
[2]黄英来,任田丽,赵鹏. VMD 与PSO 的乐器声音识别[J].哈尔滨理工大学学报,2018,(02):6.[doi:10. 15938 /j. jhust. 2018. 02. 002]
 HUANG Ying-lai,EN Tian-li,ZHAO Peng. Recognition of Instruments’Sounds Based on VMD and PSO[J].哈尔滨理工大学学报,2018,(04):6.[doi:10. 15938 /j. jhust. 2018. 02. 002]

备注/Memo

备注/Memo:
 基金项目:国家自然科学基金(11201100);黑龙江省自然科学基金(A201213)
更新日期/Last Update: 2018-10-25