Multi-class Semi-supervised Learning with the e-truncated Multinomial Probit Gaussian Process.
Simon RogersMark A. GirolamiPublished in: Gaussian Processes in Practice (2007)
Keyphrases
- gaussian process
- multi class
- semi supervised learning
- semi supervised
- pairwise
- text categorization
- gaussian processes
- labeled data
- unlabeled data
- latent variables
- approximate inference
- supervised learning
- text classification
- active learning
- probabilistic model
- binary classification
- unsupervised learning
- naive bayes
- regression model
- support vector machine
- feature selection
- machine learning
- generative model
- model selection
- em algorithm
- training data
- expectation maximization
- bayesian framework
- cost sensitive
- binary classifiers
- hyperparameters
- data sets
- multi label
- transfer learning
- maximum likelihood
- support vector
- improve the classification accuracy
- similarity measure
- markov random field
- incremental learning
- graph cuts