TuningMode (bristlecone)

TuningMode Module

Namespace: Bristlecone.Optimisation
Assembly: Bristlecone.dll
Parent Module: MonteCarlo

Functions and values

Function or value	Description
`covariance tuneInterval weighting remaining history scale` Full Usage: `covariance tuneInterval weighting remaining history scale` Parameters: tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` weighting : `float` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` scale : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>`	Tune previously observed covariance based on most recent period. Works directly with the empirical covariance of the chain history. tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` weighting : `float` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` scale : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>`
`covarianceAllTime weighting history scale` Full Usage: `covarianceAllTime weighting history scale` Parameters: weighting : `float` history : `Solution seq` scale : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>`	Tune previously observed covariance based on all time weighting : `float` history : `Solution seq` scale : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>`
`covarianceFromBounds n domain random` Full Usage: `covarianceFromBounds n domain random` Parameters: n : `int` domain : `Domain` random : `Random` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>`	Generic unit-based covariance-from-bounds n : `int` domain : `Domain` random : `Random` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>`
`covarianceFromStandardDeviations n random sigmas` Full Usage: `covarianceFromStandardDeviations n random sigmas` Parameters: n : `int` random : `Random` sigmas : `float<'u>[]` Returns: `Matrix<float<MeasureProduct<MeasureProduct<'u, 'u>, MeasureOne>>>`	Synthesises a full covariance matrix by drawing synthetic samples from marginal scales. n : `int` random : `Random` sigmas : `float<'u>[]` Returns: `Matrix<float<MeasureProduct<MeasureProduct<'u, 'u>, MeasureOne>>>`
`covarianceOnly tuneInterval weighting remaining history (arg5, arg5)` Full Usage: `covarianceOnly tuneInterval weighting remaining history (arg5, arg5)` Parameters: tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` weighting : `float` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` arg4 : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` arg5 : `'a` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>> * 'a`	tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` weighting : `float` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` arg4 : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` arg5 : `'a` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>> * 'a`
`defaultCovariance n` Full Usage: `defaultCovariance n` Parameters: n : `int` Returns: `Matrix<float>`	The starting covariance matrix for parameters in a multivariate distribution n : `int` Returns: `Matrix<float>`
`dual tuneInterval weighting remaining history (arg5, arg5)` Full Usage: `dual tuneInterval weighting remaining history (arg5, arg5)` Parameters: tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` weighting : `float` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` arg4 : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` arg5 : `float` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>> * float`	tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` weighting : `float` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` arg4 : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` arg5 : `float` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>> * float`
`dualTotalHistory tuneInterval weighting remaining history (arg5, arg5)` Full Usage: `dualTotalHistory tuneInterval weighting remaining history (arg5, arg5)` Parameters: tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` weighting : `float` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` arg4 : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` arg5 : `float` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>> * float`	tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` weighting : `float` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` arg4 : `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>>` arg5 : `float` Returns: `Matrix<float<MeasureProduct<MeasureProduct<optim-space, optim-space>, MeasureOne>>> * float`
`none arg1 arg2 factors` Full Usage: `none arg1 arg2 factors` Parameters: arg0 : `'a` arg1 : `'b` factors : `'c` Returns: `'c`	arg0 : `'a` arg1 : `'b` factors : `'c` Returns: `'c`
`samplesToMatrix samples` Full Usage: `samplesToMatrix samples` Parameters: samples : `float<'u>[][]` Returns: `Matrix<float<'u>>`	Parameters to matrix samples : `float<'u>[][]` Returns: `Matrix<float<'u>>`
`scaleFactor tuneInterval remainingIterations history scale` Full Usage: `scaleFactor tuneInterval remainingIterations history scale` Parameters: tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` remainingIterations : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` scale : `float` Returns: `float`	Modifies a scale factor tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` remainingIterations : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` scale : `float` Returns: `float`
`scaleOnly tuneInterval remaining history (arg4, arg4)` Full Usage: `scaleOnly tuneInterval remaining history (arg4, arg4)` Parameters: tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` arg3 : `'a` arg4 : `float` Returns: `'a * float`	tuneInterval : `int<MeasureProduct<iteration, MeasureOne>>` remaining : `int<MeasureProduct<iteration, MeasureOne>>` history : `Solution seq` arg3 : `'a` arg4 : `float` Returns: `'a * float`
`tuneCovariance weight recentCovariance oldCovariance` Full Usage: `tuneCovariance weight recentCovariance oldCovariance` Parameters: weight : `float` recentCovariance : `Matrix<float<'u>>` oldCovariance : `Matrix<float<'u>>` Returns: `Matrix<float<'u>>`	Tunes a covariance matrix based on recent samples. Blend two covariance matrices with a given weight. Works for any unit of measure 'u. weight : `float` recentCovariance : `Matrix<float<'u>>` oldCovariance : `Matrix<float<'u>>` Returns: `Matrix<float<'u>>`
`tuneScale scale accRate` Full Usage: `tuneScale scale accRate` Parameters: scale : `float` accRate : `float` Returns: `float`	Find an appropriate scale factor for a standard deviation and its acceptance rate. scale : `float` accRate : `float` Returns: `float`