Principled Selection of Hyperparameters in the Latent Dirichlet Allocation Model.
Clint P. GeorgeHani DossPublished in: J. Mach. Learn. Res. (2017)
Keyphrases
- hyperparameters
- model selection
- closed form
- bayesian inference
- cross validation
- random sampling
- support vector
- gaussian process
- bayesian framework
- em algorithm
- maximum likelihood
- incremental learning
- sample size
- prior information
- regularization parameter
- incomplete data
- maximum a posteriori
- noise level
- gaussian processes
- parameter space
- parameter settings
- data sets
- missing values
- higher order
- active learning
- high dimensional