Incorporating the loss function into discriminative clustering of structured outputs.
Wenliang ZhongWeike PanJames T. KwokIvor W. TsangPublished in: IEEE Trans. Neural Networks (2010)
Keyphrases
- loss function
- pairwise
- clustering method
- logistic regression
- unsupervised learning
- learning to rank
- support vector
- risk minimization
- clustering algorithm
- hinge loss
- boosting framework
- k means
- regularization term
- reproducing kernel hilbert space
- structured output
- spectral clustering
- stochastic gradient descent
- semi supervised
- empirical risk
- feature selection
- pairwise constraints
- boosting algorithms
- feature extraction
- probabilistic model
- data points
- high dimensional data
- markov random field
- convex loss functions