块主成份分析(block principal component analysis,BPCA)是一种重要的子空间学习方法,能充分利用图像矩阵的部分关联.基于L1-范数的BPCA是近年来发展起来的鲁棒降维的有效方法.本研究提出了一种新的鲁棒稀疏BPCA方法,称之为BPCAL1-S....块主成份分析(block principal component analysis,BPCA)是一种重要的子空间学习方法,能充分利用图像矩阵的部分关联.基于L1-范数的BPCA是近年来发展起来的鲁棒降维的有效方法.本研究提出了一种新的鲁棒稀疏BPCA方法,称之为BPCAL1-S.该方法相对于传统的基于L2-范数的PCA对噪声更加鲁棒.为了建立稀疏模型,优化过程中引入弹性网,联合使用Lasso与Ridge惩罚因子进行约束.提出了一种贪心算法逐个提取特征向量,对迭代过程的收敛性做了理论证明.将BPCAL1-S应用于图像分类与图像重构,实验结果验证了该方法的有效性.展开更多
With the aid of commercial finite element analysis software package ANSYS,investigations are made on the contributions of main components to stiffness of the main module for parallel machine tools,and it is found that...With the aid of commercial finite element analysis software package ANSYS,investigations are made on the contributions of main components to stiffness of the main module for parallel machine tools,and it is found that the frame is the main contributor.Then,influences of constraints,strut length and working ways of the main module have also been investigated.It can be concluded that when one of the main planes of the frame without linear drive unit is constrained,the largest whole stiffness can be acquired.And,the stiffness is much better when the main module is used in a vertical machine tool instead of a horizontal one.Finally,the principle of stiffness variation is summarized when the mobile platform reaches various positions within its working space and when various loads are applied.These achievements have provided critical instructions for the design of the main module for parallel machine tools.展开更多
文摘块主成份分析(block principal component analysis,BPCA)是一种重要的子空间学习方法,能充分利用图像矩阵的部分关联.基于L1-范数的BPCA是近年来发展起来的鲁棒降维的有效方法.本研究提出了一种新的鲁棒稀疏BPCA方法,称之为BPCAL1-S.该方法相对于传统的基于L2-范数的PCA对噪声更加鲁棒.为了建立稀疏模型,优化过程中引入弹性网,联合使用Lasso与Ridge惩罚因子进行约束.提出了一种贪心算法逐个提取特征向量,对迭代过程的收敛性做了理论证明.将BPCAL1-S应用于图像分类与图像重构,实验结果验证了该方法的有效性.
文摘With the aid of commercial finite element analysis software package ANSYS,investigations are made on the contributions of main components to stiffness of the main module for parallel machine tools,and it is found that the frame is the main contributor.Then,influences of constraints,strut length and working ways of the main module have also been investigated.It can be concluded that when one of the main planes of the frame without linear drive unit is constrained,the largest whole stiffness can be acquired.And,the stiffness is much better when the main module is used in a vertical machine tool instead of a horizontal one.Finally,the principle of stiffness variation is summarized when the mobile platform reaches various positions within its working space and when various loads are applied.These achievements have provided critical instructions for the design of the main module for parallel machine tools.