scale_logistic: Scale for a logistic data set
In mpoll/scale: The Scaleable Langevin Exact Algorithm -- Reference Implementation

Description Usage Arguments Value Author(s) References Examples

Function for running ScaLE on a logisitic data set.

1	scale_logistic(fnm = "logistic_default.RData", p.num = 2^10, t.inc = 0.01, T.extend = 0.01, run.length = 1e+05, ss.size = 10, ss.on = TRUE, seed.default = 1, data.precompute = TRUE)

`fnm`	File name output to be saved to.
`p.num`	Specify the number of particles.
`t.inc`	Specify the auxiliary mesh increment size.
`T.extend`	Specify the frequency (in diffusion time) the code should be saved.
`run.length`	Specify the total diffusion time the code has to be executed for. By default set to be 10.
`ss.size`	Specify the number of data points required within the sub-sampling construction. By default set to be 2.
`ss.on`	Specify whether to use sub-sampling for evaluation of the killing intensity. By default is set to TRUE (i.e. ON), otherwise set to FALSE (i.e. OFF).
`seed.default`	For reproducibility enter the seed for the random number generator. By default is set to 1.
`data.precompute`	Whether there exists a preloaded data file with example specification. By default is TRUE. Otherwise, specify the example you wish to run ("big.logistic.example", "rare.logistic.example" or "airline.logistic.example").

`curr.seed`	Current state of the RNG.
`time.elapse`	Total elapsed system time.
`simn`	List containing full ScaLE output (see below for list contents).
`simn$anc.times`	Vector of auxiliary mesh times
`simn$dimen`	Dimension
`simn$ess.thresh`	Percentage effective sample size threshold for conduting resampling.
`simn$log.p.wei`	Current log particle weights.
`simn$neg.wei.mech`	Mechanism by which to deal with negative particle weights (note: testing only).
`simn$p.anc.ess`	Vector of ESS calculation at auxiliary mesh times.
`simn$p.anc.neg`	Record of particles which had negative weights (note: testing only).
`simn$p.cyc.arr`	Current particle set layers with dimensions ordered by threshold passage times. Also see simn$p.pass.arr.
`simn$p.dapts`	Current data points associated with each particle.
`simn$p.idx`	Current index of each particle.
`simn$p.layer`	Current particle set ordered by next event times.
`simn$p.mat`	Current locations of particle set.
`simn$p.mu`	Initialisation point(s) of particle set.
`simn$p.num`	Number of particles
`simn$p.pass.arr`	Current particle set layers with dimensions ordered by dimension. Also see simn$p.cyc.arr.
`simn$p.path.renew`	Function used to update the particle set from one time to the next.
`simn$scale.transform`	Function used to transform points in original parametrisation space to preconditioned.
`simn$ss.phi`	Function used to evaluate true killing intensity (phi).
`simn$ss.phiC`	Function used to evaluate for a given hypercube the dominating Poisson killing intensity.
`simn$ss.size`	Subsample size.
`simn$T.fin`	Maximum diffusion time attained.
`simn$t.inc`	Auxiliary mesh increment size.
`simn$T.start`	Minimum diffusion time attained.
`simn$theta`	Vector of theta, specifying the size of the hypercube.
`simn$un.scale.transform`	Function used to transform points in preconditioned space to original parametrisation.