The Connection between Bayesian Inference and Information Theory for Model Selection, Information Gain and Experimental Design.
Sergey OladyshkinWolfgang NowakPublished in: Entropy (2019)
Keyphrases
- experimental design
- information gain
- information theory
- bayesian inference
- model selection
- feature selection
- sample size
- information theoretic
- mutual information
- hyperparameters
- text categorization
- cross validation
- probabilistic model
- statistical learning
- parameter estimation
- decision trees
- prior information
- selection criterion
- active learning
- naive bayes
- gaussian process
- support vector
- machine learning
- shannon entropy
- classification accuracy
- mixture model
- feature space
- random sampling
- feature extraction
- multi class
- text classification
- feature set
- expectation propagation
- feature subset
- generative model
- k nearest neighbor
- class imbalance
- text mining
- dimensionality reduction