Ariadne


Optimization

Nested types and modules

TypeDescription
Parameters

Type alias for kernel parameters

ModuleDescription
GradientDescent
MetropolisHastings
SquaredExp

Functions and values

Function or valueDescription
gradientDescent (...)
Signature: gradientFun:(Parameters -> Parameters) -> settings:Settings -> initialLocation:Parameters -> float []

Optimize Gaussian process hyperparameters using simple gradient descent

sampleMetropolisHastings (...)
Signature: logLikelihood:(Parameters -> float) -> transitionKernel:(Parameters -> Parameters -> float) -> proposalSampler:(Parameters -> Parameters) -> settings:Settings -> initialLocation:Parameters -> float []

Sample Gaussian process hyperparameters using MCMC with Metropolis-Hastings sampler

Fork me on GitHub