Large Margin - Minimum Classification Error Using Sum of Shifted Sigmoids as the Loss Function.
Madhavi Vedula RatnagiriBiing-Hwang JuangLawrence R. RabinerPublished in: INTERSPEECH (2011)
Keyphrases
- loss function
- minimum classification error
- discriminative training
- hinge loss
- support vector
- hidden markov models
- statistical pattern recognition
- pairwise
- probabilistic interpretation
- maximum likelihood
- support vector machine
- risk minimization
- classification error
- objective function
- maximum likelihood estimation
- automatic speech recognition
- learning algorithm
- reproducing kernel hilbert space
- gaussian mixture
- machine learning
- posterior probability
- co training
- expectation maximization