Partial correlation graphical LASSO

Partial correlation graphical LASSO
复制标题

偏相关图形LASSO

DOI:
10.1111/sjos.12675
复制
发表时间:
2023
影响因子:
1
通讯作者:
Carter J
Carter J
中科院分区:
数学4区
文献类型:
--
作者:
Carter J

文献摘要

参考文献

被引文献

相似文献

学习高斯图模型的标准似然惩罚是基于正则化精度矩阵的非对角项。这种方法,以及它们的贝叶斯对应物,对于变量的标量乘法不是不变的,除非将观察到的数据转换为单位样本方差。我们表明,这种标准化可以有很强的影响推理,并引入了一个新的家庭的惩罚偏相关的基础上。我们表明,后者,以及最大似然,L0$$ {L}_0 $$和对数罚款尺度不变。我们说明了使用这样的惩罚,偏相关图形LASSO,它设置了L1$$ {L}_1 $$的惩罚偏相关。相关的优化问题不再是凸的,而是条件凸的。我们通过模拟示例和两个真实的数据集表明,除了尺度不变之外,在推理方面可以有重要的收益。
Standard likelihood penalties to learn Gaussian graphical models are based on regularizing the off‐diagonal entries of the precision matrix. Such methods, and their Bayesian counterparts, are not invariant to scalar multiplication of the variables, unless one standardizes the observed data to unit sample variances. We show that such standardization can have a strong effect on inference and introduce a new family of penalties based on partial correlations. We show that the latter, as well as the maximum likelihood, L0$$ {L}_0 $$ and logarithmic penalties are scale invariant. We illustrate the use of one such penalty, the partial correlation graphical LASSO, which sets an L1$$ {L}_1 $$ penalty on partial correlations. The associated optimization problem is no longer convex, but is conditionally convex. We show via simulated examples and in two real datasets that, besides being scale invariant, there can be important gains in terms of inference.
DOI: 10.1016/j.ccr.2012.08.013
发表时间: 2012-11-13
期刊: Cancer cell
影响因子: 50.3
作者:
Calon A;Espinet E;Palomo-Ponce S;Tauriello DV;Iglesias M;Céspedes MV;Sevillano M;Nadal C;Jung P;Zhang XH;Byrom D;Riera A;Rossell D;Mangues R;Massagué J;Sancho E;Batlle E
通讯作者: Batlle E
惩罚高斯图模型中交叉验证的计算快速替代方案
DOI: --
发表时间: 2013
期刊:
影响因子: --
作者:
I. Vujačić;A. Abbruzzo;E. Wit
通讯作者: E. Wit
DOI: 10.4310/sii.2013.v6.n2.a8
发表时间: 2013-04-01
影响因子: 0.8
作者:
Khondker ZS;Zhu H;Chu H;Lin W;Ibrahim JG
通讯作者: Ibrahim JG
DOI: 10.1080/01621459.2018.1482755
发表时间: 2019-07-03
影响因子: 3.7
作者:
Gan, Lingrui;Narisetty, Naveen N.;Liang, Feng
通讯作者: Liang, Feng
超越套索:高斯图模型中非凸正则化的调查
DOI: --
发表时间: 2020
期刊:
影响因子: --
作者:
Williams Dr
通讯作者: Williams Dr