Subset selection for tuning of hyper-parameters in artificial neural networks.
K. K. Emre AkiTugba ErkocM. Taner EskilPublished in: ICECS (2017)
Keyphrases
- subset selection
- hyperparameters
- artificial neural networks
- model selection
- parameter settings
- cross validation
- bayesian inference
- closed form
- genetic algorithm ga
- support vector
- bayesian framework
- feature selection
- prior information
- hill climbing
- random sampling
- em algorithm
- noise level
- neural network
- variational bayes
- incremental learning
- maximum likelihood
- sample size
- incomplete data
- maximum a posteriori
- parameter optimization
- bayesian methods
- genetic algorithm
- missing values
- bayesian networks
- parameter space
- prior knowledge
- parameter values
- image reconstruction
- missing data
- generative model
- simulated annealing
- denoising
- pairwise
- search algorithm