dot-fitMCMC: Fit the MIMOSA model via MCMC

Description Usage Arguments

Description

This is an internal function that fits the MIMOSA model via MCMC. It is called from MIMOSA

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
.fitMCMC(
  data,
  inits = NULL,
  iter = 250000,
  burn = 50000,
  thin = 1,
  tune = 100,
  outfile = basename(tempfile(tmpdir = ".", fileext = ".dat")),
  alternative = "greater",
  UPPER = 0.5,
  LOWER = 0.15,
  FAST = TRUE,
  EXPRATE = 1e-04,
  pXi = c(1, 1),
  seed = 10
)

Arguments

data

a list with elements names 'n.stim' and 'n.unstim', the stimulated and unstimulated counts. Must be at least of dimension 2.

inits

the initialization parameters for the MCMC routine. Can be initialized from MDMix with initonly=TRUE.

iter

the number of Mote Carlo iterations

burn

the number of burn-in iterations

thin

The thinning interval

tune

the number of iterations used for tuning the step size

outfile

the output file name

alternative

either 'greater' or 'not equal' for fitting the one-sided or two-sided MIMOSA model, respectively.

UPPER

tuning parameter for the upper bound on the acceptance ratio of each paramter

LOWER

tuning parmeter for the lower bound on the acceptance ratio of each paramter

FAST

TRUE,FALSE. Use the heuristic (FAST=TRUE) for fitting a one-sided model rather than recomputing the normalization constant via MCMC for each step. @importFrom coda mcmc

EXPRATE

the mean of the prior distribution for the model hyperparameters.

pXi

is the prior on the w, beta(1,1) by default).

seed

numeric random seed @rdname fitMCMC @name .fitMCMC @importFrom data.table fread


MIMOSA documentation built on Nov. 12, 2020, 2:02 a.m.