Derivation of Update Rules for Convolutive NMF Based on Squared Euclidean Distance, KL Divergence, and IS Divergence.
Hiroki TanjiRyo TanakaKyohei TabataYoshito IsekiTakahiro MurakamiYoshihisa IshidaPublished in: IEICE Trans. Fundam. Electron. Commun. Comput. Sci. (2014)
Keyphrases
- kl divergence
- bregman divergences
- mahalanobis distance
- kullback leibler divergence
- information theoretic
- gaussian distribution
- exponential family
- gaussian mixture
- posterior distribution
- mutual information
- dissimilarity measure
- probabilistic latent semantic analysis
- noise model
- missing values
- face recognition
- probability density
- negative matrix factorization
- probability density function
- closed form
- sample size