证明置换理想上的正则矩阵可以通过满矩阵对角化。
We show that every regular matrix over exchange ideals admits a diagonal reduction by full matrices.
提出一种带有正则约束的非负矩阵分解算法(RCNMF)。
An algorithm with regularization constrains for nonnegative matrix factorization (RCNMF) is proposed.
其基本思想是利用验前模型信息确定正则化矩阵,利用验后观测信息确定自适应因子。
The regularization matrix will be chosen by prior information of the model parameters and the adaptive factor, however, will be determined by posterior information.
应用推荐