Fast bregman divergence NMF using taylor expansion and coordinate descent.
Liangda LiGuy LebanonHaesun ParkPublished in: KDD (2012)
Keyphrases
- bregman divergences
- nonnegative matrix factorization
- cost sensitive
- information theoretic
- maximum entropy
- learning theory
- negative matrix factorization
- mahalanobis distance
- exponential family
- loss function
- nearest neighbor
- theoretical guarantees
- data dependent
- special case
- maximum likelihood
- kl divergence
- learning algorithm
- spectral clustering
- mutual information
- matrix factorization
- least squares
- covariance matrix
- boosting algorithms
- sparse representation
- naive bayes
- active learning
- feature space
- bayesian networks
- similarity measure