A genetic mixed-integer optimization of neural network hyper-parameters.
Kyle SpurlockHeba ElgazzarPublished in: J. Supercomput. (2022)
Keyphrases
- mixed integer
- hyperparameters
- neural network
- model selection
- cross validation
- genetic algorithm
- linear program
- random sampling
- lot sizing
- feasible solution
- convex hull
- bayesian framework
- em algorithm
- support vector
- bayesian inference
- closed form
- optimal solution
- data sets
- optimization problems
- continuous variables
- incomplete data
- sample size
- noise level
- incremental learning
- maximum a posteriori
- low dimensional
- maximum likelihood
- probability distribution
- expectation maximization
- prior knowledge
- machine learning
- posterior distribution
- computational complexity
- tabu search
- missing values