Strong Coresets for Hard and Soft Bregman Clustering with Applications to Exponential Family Mixtures.
Mario LucicOlivier BachemAndreas KrausePublished in: CoRR (2015)
Keyphrases
- exponential family
- bregman divergences
- mixture model
- density estimation
- maximum likelihood
- clustering algorithm
- information theoretic
- log likelihood
- closed form
- missing values
- kl divergence
- unsupervised learning
- k means
- graphical models
- statistical models
- maximum entropy
- outlier detection
- probability density function
- expectation maximization
- nonnegative matrix factorization
- order statistics
- cost sensitive
- bayesian framework
- nearest neighbor
- mahalanobis distance
- loss function