Directionally Paired Principal Component Analysis for Bivariate Estimation Problems.

Directionally Paired Principal Component Analysis for Bivariate Estimation Problems.
复制标题

DOI:
10.1109/icpr48806.2021.9412245
复制
发表时间:
2021-01
期刊:
Proceedings of the ... IAPR International Conference on Pattern Recognition. International Conference on Pattern Recognition
影响因子:
--
通讯作者:
Yezzi A
Yezzi A
中科院分区:
其他
文献类型:
--
作者:
Fan Y;Dahiya N;Bignardi S;Sandhu R;Yezzi A

文献摘要

参考文献

相似文献

提出了一种新的线性降维模型--方向配对主成分分析(DP-PCA),用于估计耦合但部分可观测的变量集。与偏最小二乘方法(例如偏最小二乘回归和典型相关分析)最大化两个数据集之间的相关性/协方差不同,我们的DP-PCA直接有条件或无条件地最小化可观测部分和不可观测部分的重建和预测误差。通过在合成高斯数据、多目标回归数据集和单通道图像数据集上的数据重建和预测实验,我们对相关的线性交叉分解方法进行了比较和评估。结果表明,当只允许一对碱基时,条件DP-PCA在可观测部分和总体变量集上的重建误差最小,而无条件DP-PCA在不可观测部分达到最小的预测误差。当允许额外的预算用于可观测部分的主成分分析时,可以使用组合方法得到最优解:可观测部分的标准主成分分析和不可观测部分的无条件DP-PCA。
We propose Directionally Paired Principal Component Analysis (DP-PCA), a novel linear dimension-reduction model for estimating coupled yet partially observable variable sets. Unlike partial least squares methods (e.g., partial least squares regression and canonical correlation analysis) that maximize correlation/covariance between the two datasets, our DP-PCA directly minimizes, either conditionally or unconditionally, the reconstruction and prediction errors for the observable and unobservable part, respectively. We demonstrate the optimality of the proposed DP-PCA approach, we compare and evaluate relevant linear cross-decomposition methods with data reconstruction and prediction experiments on synthetic Gaussian data, multi-target regression datasets, and a single-channel image dataset. Results show that when only a single pair of bases is allowed, the conditional DP-PCA achieves the lowest reconstruction error on the observable part and the total variable sets as a whole; meanwhile, the unconditional DP-PCA reaches the lowest prediction errors on the unobservable part. When an extra budget is allowed for the observable part’s PCA basis, one can reach an optimal solution using a combined method: standard PCA for the observable part and unconditional DP-PCA for the unobservable part.
DOI: 10.1002/joc.3370140404
发表时间: 1994-05-01
期刊: INTERNATIONAL JOURNAL OF CLIMATOLOGY
影响因子: --
作者:
COOK, ER;BRIFFA, KR;JONES, PD
通讯作者: JONES, PD
DOI: 10.1093/biomet/28.3-4.321
发表时间: 1936-12-01
期刊: BIOMETRIKA
影响因子: 2.7
作者:
Hotelling, H
通讯作者: Hotelling, H
DOI: 10.1080/14786440109462720
发表时间: 1901-07-01
影响因子: 1.6
作者:
Pearson, Karl
通讯作者: Pearson, Karl
DOI: 10.1002/aic.690370209
发表时间: 1991-02-01
期刊: AICHE JOURNAL
影响因子: 3.7
作者:
KRAMER, MA
通讯作者: KRAMER, MA
DOI: 10.1007/s10994-016-5546-z
发表时间: 2016-07-01
期刊: MACHINE LEARNING
影响因子: 7.5
作者:
Spyromitros-Xioufis, Eleftherios;Tsoumakas, Grigorios;Vlahavas, Ioannis
通讯作者: Vlahavas, Ioannis