BriefGPT.xyz
Nov, 2016
鲁棒多模型拟合的非负矩阵欠拟合
Nonnegative Matrix Underapproximation for Robust Multiple Model Fitting
HTML
PDF
Mariano Tepper, Guillermo Sapiro
TL;DR
本文介绍了一种高效算法,用于解决非负矩阵欠逼近(NMU)问题,提出NMU结果与传统NMF相比具有额外的稀疏性和基于部分行为,解释了独特的数据特征。通过应用到气候数据分析和多参数模型拟合,证明了该方法在NMU计算效率上的优越性和实用性。
Abstract
In this work, we introduce the first algorithm to truly address the
nonnegative matrix underapproximation
(NMU) problem, i.e.,
nonnegative matrix factorization
(NMF) with an additional underapproximation constrai
→