Krijnen Wim P, Dijkstra Theo K, Stegeman Alwin
Psychometrika. 2008 Sep;73(3):431-439. doi: 10.1007/s11336-008-9056-1. Epub 2008 Jan 29.
The CANDECOMP/PARAFAC (CP) model decomposes a three-way array into a prespecified number of R factors and a residual array by minimizing the sum of squares of the latter. It is well known that an optimal solution for CP need not exist. We show that if an optimal CP solution does not exist, then any sequence of CP factors monotonically decreasing the CP criterion value to its infimum will exhibit the features of a so-called "degeneracy". That is, the parameter matrices become nearly rank deficient and the Euclidean norm of some factors tends to infinity. We also show that the CP criterion function does attain its infimum if one of the parameter matrices is constrained to be column-wise orthonormal.
CANDECOMP/PARAFAC(CP)模型通过最小化残差阵列的平方和,将一个三维阵列分解为预先指定数量的R个因子和一个残差阵列。众所周知,CP的最优解不一定存在。我们证明,如果不存在CP最优解,那么任何将CP准则值单调递减至其下确界的CP因子序列都会呈现出所谓“退化”的特征。也就是说,参数矩阵几乎变得秩亏缺,并且某些因子的欧几里得范数趋于无穷大。我们还证明,如果其中一个参数矩阵被约束为列正交归一,则CP准则函数确实会达到其下确界。