Unsupervised learning of parsimonious mixtures on large spaces with integrated feature and component selection

Keyphrases

• Bayesian Information Criterion 100%
• Unsupervised Learning 100%
• Component Selection 100%
• Large Space 100%
• Integrated Components 100%
• Feature Space 50%
• Multinomial 50%
• Classification Performance 50%
• Mixture Model 50%
• Number of Components 50%
• Model Complexity 50%
• Feature Selection 50%
• Common Model 50%
• Misclassification Error 50%
• Text Data 50%
• Mixing Parameters 50%
• Number of Subjects 50%
• Unsupervised Method 50%
• Generalized Expectation Maximization 50%
• Parsimonious Models 50%
• Naïve Bayes Classifier 50%
• Feature Subset 50%
• Data Likelihood 50%

Computer Science

• Unsupervised Learning 100%
• Information Criterion 100%
• Feature Selection 50%
• Model Complexity 50%
• Classification Error 50%
• Feature Space 50%
• Bayes Classifier 50%
• Labeled Example 50%
• Message Length 50%
• Generalized Expectation 50%
• Unsupervised Method 50%
• Reduce Complexity 50%