Log-likelihood method to select initial values of multichannel non-negative matrix factorization
Log-likelihood method to select initial values of multichannel non-negative matrix factorization
复制标题
对数似然法选择多通道非负矩阵分解初始值
DOI：
10.1109/icce-tw.2015.7216813
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
K
中科院分区：
文献类型：
--
作者：
Yoshiyama;F.;Uenohara;S.;Nishijima;K.;Hioka;Y.;& Furuya;K
非负矩阵分解（NMF）的多通道扩展将源的空间特性与每个NMF基相关联。提出了一种基于对数似然的多通道非负矩阵分解（MNMF）初始值选择方法，以减小源分离性能的变化。实验结果表明，选择具有高对数似然的初始值可以提高MNMF的源分离性能。
A multichannel extension of non-negative matrix factorization (NMF) associates the spatial property of the sources with each of the NMF bases. An initial-value selection method based on log-likelihood for multichannel non-negative matrix factorization (MNMF) is introduced to reduce the variation of the source separation performance. Experimental results showed selecting initial values that provide high log-likelihood would improve the source separation performance of MNMF depending on the sources.