plot.hyper: Plot DMC Samples at the Hyper level

In TasCL/ggdmc: Dynamic Model of Choice with Parallel Computation, and C++ Capabilities

Description

Plot trace and probability desntiy. This method provides a switch to choose plotting at the data or hyper level. The samples must be an object fit by a hierarchical model.

Usage

## S3 method for class 'hyper'
plot(x, y = NULL, hyper = FALSE, start = 1, end = NA,
  save.ll = FALSE, main.pll = NULL, pll.chain = FALSE,
  pll.together = TRUE, pll.barplot = FALSE, only.prior = FALSE,
  only.like = FALSE, subject = 1, smooth = FALSE, density = FALSE,
  save.dat = FALSE, p.prior = NULL, natural = TRUE, trans = NA,
  xlim = NA, chain1 = TRUE, ...)

Arguments

x	run.dmc or samples.dmc generated model samples
y	default NULL. No function. Just to make it compatible to plot
hyper	a boolean switch to draw hyper-parameter. By default it is off.
start	instruct the function to plot starting from which iteration. This indicates how many burn-in interations one requests. For example, start=101, indicates 100 burn-in interations.
end	instruct the function to plot ending at a certain iteration
save.ll	a boolean switch to tell the function to save the mean log-likelihood. This option does not work in DMC's plot.dmc, too.
main.pll	a string as the title for the boxplot. Default is NULL
pll.chain	a boolean switch to plot posterior log likelihoood
pll.together	a boolean switch to plot the posterior log-likelihood chains all together in one canvar
pll.barplot	a boolean switch to plot the means of posterior log-likelihood of all chains as a barplot. By default, it is off.
only.prior	only.prior and only.like pick out one or other component of pll.
only.like	a boolean switch.
subject	indicate which subject to plot in a multi-subject list samples
smooth	default FALSE
density	plot probability density too?
save.dat	whether save the internal data table out for polish plots
p.prior	prior distribution setting. necessary for plot.prior to work
natural	additional argument for plot.prior
trans	additional argument for plot.prior
xlim	additional argument for plot.prior
chain1	plot all chains or just chain1 as coda does
...	other arguments

Examples

m1 <- model.dmc(
p.map     = list(a="1",v="1",z="1",d="1",sz="1",sv="1",t0="1",st0="1"),
match.map = list(M=list(s1="r1", s2="r2")),
factors   = list(S=c("s1", "s2")),
constants = c(st0=0, d=0),
responses = c("r1","r2"),
type      = "rd")

## Population distribution
pop.prior <- prior.p.dmc(
  dists = rep("tnorm", 6),
  p1    = c(a=2,   v=2.5, z=0.5, sz=0.3, sv=1,  t0=0.3),
  p2    = c(a=0.5, v=.5,  z=0.1, sz=0.1, sv=.3, t0=0.05),
  lower = c(0,-5, 0, 0, 0, 0),
  upper = c(5, 7, 2, 2, 2, 2))

dat <- h.simulate.dmc(m1, p.prior=pop.prior, n=50, ns=4)
mdi <- data.model.dmc(dat, m1)
p.prior  <- prior.p.dmc(
  dists = rep("tnorm", 6),
  p1    = c(a=2,   v=2.5, z=0.5, sz=0.3, sv=1,  t0=0.3),
  p2    = c(a=0.5, v=.5,  z=0.1, sz=0.1, sv=.3, t0=0.05) * 5,
  lower = c(0,-5, 0, 0, 0, 0),
  upper = c(5, 7, 2, 2, 2, 2))

## Fixed-effect model
samplesInit <- h.samples.dmc(nmc=100, p.prior=p.prior, data=mdi, thin=1)
samples0    <- h.run.dmc(samples=samplesInit, report=20)

## Make a hyper-prior list
mu.prior <- prior.p.dmc(
  dists = rep("tnorm", 6),
  p1    = c(a=2,   v=2.5, z=0.5, sz=0.3, sv=1,  t0=0.3),
  p2    = c(a=0.5, v=.5,  z=0.1, sz=0.1, sv=.3, t0=0.05) * 5,
  lower = c(0,-5, 0, 0, 0, 0),
  upper = c(5, 7, 2, 2, 2, 2))

sigma.prior <- prior.p.dmc(
  dists = rep("beta", 6),
  p1    = c(a=1, v=1, z=1, sz=1, sv=1, t0=1),
  p2    = c(1,1,1,1,1,1),
  upper = c(2,2,2,2,2,2))

pp.prior <- list(mu.prior, sigma.prior)

## Random-effect model
hsamplesInit <- h.samples.dmc(nmc=50, p.prior=p.prior, pp.prior=pp.prior,
  data=mdi, thin=1)
hsamples0 <- h.run.dmc(samples=hsamplesInit)

## Windows Testing errors
## plot(hsamples0) ## Only first participant
## plot(hsamples0, hyper=TRUE) ## Group-level parameters
## plot(hsamples0, hyper=TRUE, density=TRUE) ## Trace and density plots

## plot(hsamples0, p.prior=p.prior) ## plot prior and posterior
## plot(hsamples0, p.prior=pp.prior, hyper=TRUE) ## hyper-level

## plot(hsamples0, hyper=TRUE, start=101) ## starting from 101th iteration
## plot(hsamples0, hyper=TRUE, density=TRUE, start=101)

## Plot posterior likelihood
## plot(hsamples0, hyper=TRUE, pll.chain=TRUE, start=21)

## Save plot data for modifying
## D <- plot(hsamples0, hyper=TRUE, save.dat=TRUE)
## head(D)
## Source: local data frame [6 x 4]
##   Iteration  Chain Parameter    value
##       (int) (fctr)    (fctr)    (dbl)
## 1         1      1      a.h1 2.184504
## 2         2      1      a.h1 2.184504
## 3         3      1      a.h1 2.184504
## 4         4      1      a.h1 2.501221
## 5         5      1      a.h1 2.501221
## 6         6      1      a.h1 2.501221

TasCL/ggdmc documentation built on May 9, 2019, 4:19 p.m.