Struct StochasticNLL 

pub struct StochasticNLL {
    pub nll: NLL,
    /* private fields */
}

A stochastic NLL term.

While a regular NLL will operate over the entire dataset, this term will only operate over a random subset of the data, determined by the batch_size parameter. This will make the objective function faster to evaluate at the cost of adding random noise to the likelihood.

Fields

nll: NLL

A handle to the original NLL term.

Implementations

impl StochasticNLL

pub fn minimize( &self, settings: MinimizationSettings<Self>, ) -> Result<MinimizationSummary, LadduError>

Minimize the StochasticNLL with the algorithm given in the MinimizationSettings.

Errors

This method may return an error if there was any problem constructing thread pools or evaluating the underlying model.

pub fn mcmc( &self, settings: MCMCSettings<Self>, ) -> Result<MCMCSummary, LadduError>

Run an MCMC sampling algorithm over the StochasticNLL with the given MCMCSettings.

Errors

This method may return an error if there was any problem constructing thread pools or evaluating the underlying model.

impl StochasticNLL

pub fn new(nll: NLL, batch_size: usize, seed: Option<usize>) -> Self

Generate a new StochasticNLL with the given NLL, batch size, and optional random seed

See Also

NLL::to_stochastic

pub fn resample(&self)

Resample the batch indices used in evaluation

Trait Implementations

impl Clone for StochasticNLL

fn clone(&self) -> StochasticNLL

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl CostFunction<MaybeThreadPool, LadduError> for StochasticNLL

fn evaluate( &self, parameters: &DVector<Float>, args: &MaybeThreadPool, ) -> Result<Float, LadduError>

The evaluation of the function at a point x with the given arguments/user data.

impl Gradient<MaybeThreadPool, LadduError> for StochasticNLL

fn gradient( &self, parameters: &DVector<Float>, args: &MaybeThreadPool, ) -> Result<DVector<Float>, LadduError>

The evaluation of the gradient at a point x with the given arguments/user data.

fn hessian( &self, x: &Matrix<f64, Dyn, Const<1>, VecStorage<f64, Dyn, Const<1>>>, args: &U, ) -> Result<Matrix<f64, Dyn, Dyn, VecStorage<f64, Dyn, Dyn>>, E>

The evaluation of the hessian at a point x with the given arguments/user data.

impl LikelihoodTerm for StochasticNLL

fn parameters(&self) -> Vec<String>

The list of names of the input parameters for LikelihoodTerm::evaluate.

fn evaluate(&self, parameters: &[Float]) -> Float

Evaluate the term given some input parameters.

fn evaluate_gradient(&self, parameters: &[Float]) -> DVector<Float>

Evaluate the gradient of the term given some input parameters.

fn update(&self)

A method called every step of any minimization/MCMC algorithm.

impl LogDensity<MaybeThreadPool, LadduError> for StochasticNLL

fn log_density( &self, parameters: &DVector<Float>, args: &MaybeThreadPool, ) -> Result<Float, LadduError>

The log of the evaluation of the density function at a point x with the given arguments/user data.