The influence of hyper-parameters in the infinite relational model.
Kristoffer Jon AlbersMorten MørupMikkel N. SchmidtPublished in: MLSP (2016)
Keyphrases
- hyperparameters
- model selection
- cross validation
- bayesian inference
- closed form
- random sampling
- bayesian framework
- prior information
- gaussian process
- em algorithm
- support vector
- posterior distribution
- maximum likelihood
- noise level
- bayesian methods
- incremental learning
- incomplete data
- sample size
- variational bayes
- maximum a posteriori
- active learning
- parameter space
- parameter settings
- multiresolution
- feature selection
- data mining
- data sets