Universal Clustering with Family of Power Loss Functions in Probabilistic Space.
Vladimir NikulinPublished in: IDEAL (2005)
Keyphrases
- loss function
- reproducing kernel hilbert space
- learning to rank
- bayes rule
- pairwise
- loss minimization
- clustering method
- clustering algorithm
- support vector
- k means
- squared error
- probabilistic model
- information theoretic
- risk minimization
- pairwise constraints
- information retrieval
- hinge loss
- convex loss functions
- data clustering
- spectral clustering
- bregman divergences
- document clustering
- vector space
- distance metric
- high dimensional data
- labeled data
- active learning
- training data
- update rules
- learning algorithm
- square loss
- machine learning