Lasso penalized model selection criteria for high-dimensional multivariate linear regression analysis
Lasso penalized model selection criteria for high-dimensional multivariate linear regression analysis
复制标题
DOI：
10.1016/j.jmva.2014.08.002
复制
发表时间：
2014-11
期刊：
影响因子：
--
通讯作者：
Shota Katayama;S. Imori
中科院分区：
文献类型：
--
作者：
Shota Katayama;S. Imori
本文提出了在高维多元线性回归分析中识别相关预测因子的两个模型选择准则。建议的标准是基于Lasso型惩罚似然函数，允许高维。在多个响应维数趋于无穷大，而候选模型的最大样本量具有较小的样本量阶数的渐近框架下，证明了所提出的准则具有模型选择的一致性，即它们能渐近地选出真实模型.仿真研究表明，当多个响应的维数较大时，所提出的准则优于现有准则。
This paper proposes two model selection criteria for identifying relevant predictors in the high-dimensional multivariate linear regression analysis. The proposed criteria are based on a Lasso type penalized likelihood function to allow the high-dimensionality. Under the asymptotic framework that the dimension of multiple responses goes to infinity while the maximum size of candidate models has smaller order of the sample size, it is shown that the proposed criteria have the model selection consistency, that is, they can asymptotically pick out the true model. Simulation studies show that the proposed criteria outperform existing criteria when the dimension of multiple responses is large.