基于矩阵变量分布的分类。
Classification with the matrix-variate- distribution.
作者信息
Thompson Geoffrey Z, Maitra Ranjan, Meeker William Q, Bastawros Ashraf F
机构信息
Iowa State University, Ames, Iowa, USA.
出版信息
J Comput Graph Stat. 2020;29(3):668-674. doi: 10.1080/10618600.2019.1696208. Epub 2020 Jan 22.
Abstract
Matrix-variate distributions can intuitively model the dependence structure of matrix-valued observations that arise in applications with multivariate time series, spatio-temporal or repeated measures. This paper develops an Expectation-Maximization algorithm for discriminant analysis and classification with matrix-variate -distributions. The methodology shows promise on simulated datasets or when applied to the forensic matching of fractured surfaces or the classification of functional Magnetic Resonance, satellite or hand gestures images.
摘要
矩阵变量分布可以直观地对矩阵值观测的依赖结构进行建模，这些观测值出现在多元时间序列、时空或重复测量的应用中。本文开发了一种用于判别分析和分类的期望最大化算法，该算法基于矩阵变量分布。该方法在模拟数据集上或应用于断裂表面的法医匹配、功能磁共振、卫星或手势图像的分类时显示出前景。
相似文献
1
Classification with the matrix-variate- distribution.基于矩阵变量分布的分类。
J Comput Graph Stat. 2020;29(3):668-674. doi: 10.1080/10618600.2019.1696208. Epub 2020 Jan 22.
2
Quantitative matching of forensic evidence fragments utilizing 3D microscopy analysis of fracture surface replicas.利用断裂面复制品的 3D 显微镜分析对法医证据碎片进行定量匹配。
J Forensic Sci. 2022 May;67(3):899-910. doi: 10.1111/1556-4029.15012. Epub 2022 Mar 7.
3
A theoretical framework for Landsat data modeling based on the matrix variate mean-mixture of normal model.基于矩阵变量均值-混合正态模型的 Landsat 数据建模理论框架。
PLoS One. 2020 Apr 9;15(4):e0230773. doi: 10.1371/journal.pone.0230773. eCollection 2020.
4
Robust Bilinear Probabilistic PCA Using a Matrix Variate t Distribution.使用矩阵变量t分布的稳健双线性概率主成分分析
IEEE Trans Neural Netw Learn Syst. 2023 Dec;34(12):10683-10697. doi: 10.1109/TNNLS.2022.3170797. Epub 2023 Nov 30.
5
A Bayes optimal matrix-variate LDA for extraction of spatio-spectral features from EEG signals.一种用于从脑电图信号中提取时空频谱特征的贝叶斯最优矩阵变量线性判别分析方法。
Annu Int Conf IEEE Eng Med Biol Soc. 2012;2012:3955-8. doi: 10.1109/EMBC.2012.6346832.
6
Block Sparse Variational Bayes Regression Using Matrix Variate Distributions With Application to SSVEP Detection.使用矩阵变量分布的块稀疏变分贝叶斯回归及其在稳态视觉诱发电位检测中的应用
IEEE Trans Neural Netw Learn Syst. 2022 Jan;33(1):351-365. doi: 10.1109/TNNLS.2020.3027773. Epub 2022 Jan 5.
7
Model-based learning using a mixture of mixtures of Gaussian and uniform distributions.基于模型的学习，使用混合高斯和均匀分布的混合物。
IEEE Trans Pattern Anal Mach Intell. 2012 Apr;34(4):814-7. doi: 10.1109/TPAMI.2011.199.
8
Parsimonious hidden Markov models for matrix-variate longitudinal data.用于矩阵变量纵向数据的简约隐马尔可夫模型。
Stat Comput. 2022;32(3):53. doi: 10.1007/s11222-022-10107-0. Epub 2022 Jun 15.
9
Modal clustering of matrix-variate data.矩阵变量数据的模态聚类
Adv Data Anal Classif. 2023;17(2):323-345. doi: 10.1007/s11634-022-00501-x. Epub 2022 May 5.
10
Extracting urban impervious surfaces from Sentinel-2 and Landsat-8 satellite data for urban planning and environmental management.从 Sentinel-2 和 Landsat-8 卫星数据中提取城市不透水面，用于城市规划和环境管理。
Environ Sci Pollut Res Int. 2021 Feb;28(6):6572-6586. doi: 10.1007/s11356-020-11007-4. Epub 2020 Oct 1.
引用本文的文献
1
Construction of a poor prognosis prediction and visualization system for intracranial aneurysm endovascular intervention treatment based on an improved machine learning model.基于改进机器学习模型的颅内动脉瘤血管内介入治疗不良预后预测与可视化系统构建
Front Neurol. 2025 Jan 8;15:1482119. doi: 10.3389/fneur.2024.1482119. eCollection 2024.
2
Quantitative matching of forensic evidence fragments using fracture surface topography and statistical learning.利用断口表面形貌和统计学习对法医证据碎片进行定量匹配
Nat Commun. 2024 Sep 8;15(1):7852. doi: 10.1038/s41467-024-51594-1.
3
Quantitative matching of forensic evidence fragments utilizing 3D microscopy analysis of fracture surface replicas.利用断裂面复制品的 3D 显微镜分析对法医证据碎片进行定量匹配。
J Forensic Sci. 2022 May;67(3):899-910. doi: 10.1111/1556-4029.15012. Epub 2022 Mar 7.
4
A Matrix-Variate Model for Networks.一种用于网络的矩阵变量模型。
Front Artif Intell. 2021 May 13;4:674166. doi: 10.3389/frai.2021.674166. eCollection 2021.
5
Simultaneous differential network analysis and classification for matrix-variate data with application to brain connectivity.同时进行矩阵变量数据的差异网络分析和分类及其在脑连接中的应用。
Biostatistics. 2022 Jul 18;23(3):967-989. doi: 10.1093/biostatistics/kxab007.
本文引用的文献
1
Fast Adaptive Smoothing and Thresholding for Improved Activation Detection in Low-Signal fMRI.快速自适应平滑和阈值处理，提高低信号 fMRI 中的激活检测。
IEEE Trans Med Imaging. 2019 Dec;38(12):2821-2828. doi: 10.1109/TMI.2019.2915052. Epub 2019 May 7.
2
A re-defined and generalized percent-overlap-of-activation measure for studies of fMRI reproducibility and its use in identifying outlier activation maps.一种重新定义和广义的 fMRI 可重复性研究的激活重叠百分比度量方法及其在识别异常激活图中的应用。
Neuroimage. 2010 Mar;50(1):124-35. doi: 10.1016/j.neuroimage.2009.11.070. Epub 2009 Dec 3.
3
Assessing certainty of activation or inactivation in test-retest fMRI studies.在重测功能磁共振成像研究中评估激活或失活的确定性。
Neuroimage. 2009 Aug 1;47(1):88-97. doi: 10.1016/j.neuroimage.2009.03.073. Epub 2009 Apr 8.
4
Uncorrelated multilinear discriminant analysis with regularization and aggregation for tensor object recognition.用于张量对象识别的带正则化和聚合的不相关多线性判别分析
IEEE Trans Neural Netw. 2009 Jan;20(1):103-23. doi: 10.1109/TNN.2008.2004625. Epub 2008 Dec 12.
5
Multilinear discriminant analysis for face recognition.用于人脸识别的多线性判别分析。
IEEE Trans Image Process. 2007 Jan;16(1):212-20. doi: 10.1109/tip.2006.884929.
6
Test-retest reliability estimation of functional MRI data.功能磁共振成像数据的重测信度估计
Magn Reson Med. 2002 Jul;48(1):62-70. doi: 10.1002/mrm.10191.
Zotero 插件