A three–step algorithm for CANDECOMP/PARAFAC analysis of large data sets with multicollinearity
A three–step algorithm for CANDECOMP/PARAFAC analysis of large data sets with multicollinearity
复制标题
具有多重共线性的大数据集 CANDECOMP/PARAFAC 分析的三步算法
DOI：
10.1002/(sici)1099-128x(199805/06)12:3
复制
发表时间：
1998
影响因子：
2.4
通讯作者：
H. Kiers
中科院分区：
文献类型：
--
作者：
H. Kiers
通过标准的交替最小二乘算法拟合CANDECOMP/PARAFAC模型通常需要非常多的迭代。一个恰当的例子是分析具有轻微到严重多重共线性的数据。此外，如果数据量很大，一个CANDECOMP/PARAFAC解决方案的计算是非常耗时的。本文通过将数据压缩的思想与压缩数据数组的一种正则化形式相结合，描述了一个比普通CANDECOMP/PARAFAC算法更有效的三步程序。©1998 John Wiley & Sons, Ltd
Fitting the CANDECOMP/PARAFAC model by the standard alternating least squares algorithm often requires very many iterations. One case in point is that of analysing data with mild to severe multicollinearity. If, in addition, the size of the data is large, the computation of one CANDECOMP/PARAFAC solution is very time‐consuming. The present paper describes a three‐step procedure which is much more efficient than the ordinary CANDECOMP/PARAFAC algorithm, by combining the idea of data compression with a form of regularization of the compressed data array. © 1998 John Wiley & Sons, Ltd.