期刊文献+

利用交替方向法求解H权重最近相关系数矩阵问题

ADM for H-weighted nearest correlation matrix problem
下载PDF
导出
摘要 由于计算H权重的半正定矩阵锥投影比较困难,目前求解带有H权重的最近相关系数矩阵问题的方法很少且比较复杂.考虑用交替方向法求解该问题,每次迭代只需求解一个有显式解的二次规划问题和一个不带权重的半正定矩阵锥投影,计算简单,易于实现.为提高计算速度,还考虑了改进的交替方向法.此外,通过数值实验对交替方向法与现有方法进行了比较,说明了交替方向法对解决带有H权重的最近相关系数矩阵问题的有效性. There are few methods for solving H-weighted nearest correlation matrix (HNCM) problem in the literatures.The main reason is that it is difficult to obtain a computable formula for the proj ection of a matrix onto the positive semidefinite matrix cone under the H-weight.The alternating direction method (ADM)is applied for solving HNCM problems.At each iteration,a quadratic program with closed-form solution and a proj ection on positive semidefinite matrix cone without weight are calculated.The method is very simple and easy to be implemented.To speed up the calculation,a modified ADM is also studied.The numerical results show that compared with the existing methods ADM is very efficient for solving HNCM problem.
出处 《大连理工大学学报》 EI CAS CSCD 北大核心 2015年第5期553-558,共6页 Journal of Dalian University of Technology
基金 国家自然科学基金资助项目(11101064)
关键词 交替方向法 最近相关系数矩阵 半正定矩阵锥 alternating direction method the nearest correlation matrix positive semidefinite matrix cone
  • 相关文献

参考文献10

  • 1Higham N J. Computing the nearest correlationmatrix - a problem from finance [J]. IMA Journalof Numerical Analysis,2002,22(3) :329-343.
  • 2Borsdorf R,Higham N J. A preconditioned Newtonalgorithm for the nearest correlation matrix [J ].IMA Journal of Numerical Analysis,2010,30(1);94-107.
  • 3QI Hou-duo,SUN De-feng. A quadraticallyconvergent Newton method for computing thenearest correlation matrix [J]. SIAM Journal onMatrix Analysis and Applications. 2006,28(2) :360-385.
  • 4Toh K C. An inexact primal-dual path followingalgorithm for convex quadratic SDP [ J ].Mathematical Programming, 2008, 112 ( 1 ): 221-254.
  • 5Malick J. A dual approach to semidefinite least-squares problems [J ]. SIAM Journal on MatrixAnalysis and Applications. 2004, 26(1):272-284.
  • 6QI Hou-duo, SUN De-feng. An augmentedLagrangian dual approach for the H-weightednearest correlation matrix problem [J]. IMAJournal of Numerical Analysis, 2011,31(2): 491-511.
  • 7Boyd S,Parikh N, Chu E,et al. Distributedoptimization and statistical learning via thealternating direction method of multipliers [J].Foundations and Trends in Machine Learning, 2010,3(1):1-122.
  • 8YECai-hong,YUAN Xiao-ming. A descent methodfor structured monotone variational inequalities [J].Optimization Methods and Software, 2007,22 (2);329-338.
  • 9He B, Xu M, Yuan X. Solving large-scale leastsquares semidefinite programming by alternatingdirection methods [J]. SIAM Journal on MatrixAnalysis and Applications, 2011,32(1) :136-152.
  • 10LI Qing-na,QI Hou-duo. A sequential semismoothNewton method for the nearest low-rank correlationmatrix problem [J]. SIAM Journal on Optimization,2011, 21(4):1641-1666.

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部