plot.timedeppar {timedeppar}

Plot results of time-dependent parameter estimation

Description

This function plot Markov chains and marginal densities of constant parameters, distributions of time dependent parameters, and Markov chains and marginal densities of time-dependent parameters at selected points in time.

Usage

## S3 method for class 'timedeppar'
plot(
  x,
  type = c("traces", "marginals", "summary", "pairs", "time-series", "accept"),
  chains.at = numeric(0),
  labels = NA,
  units = NA,
  prob.band = 0.9,
  max.diag.plots = 100,
  xlim.ts = numeric(0),
  n.burnin = 0,
  nrow = 4,
  nrow.constpar = NA,
  nrow.timedeppar = NA,
  nrow.diagnostics = NA,
  ...
)

Arguments

x

results from the function infer.timedeppar of class timedeppar or list of such results for comparing multiple chains (in the latter case you have to call explicitly plot.timedeppar rather than being able to do the generic call plot as the list of results is not an object of class timedeppar).

type

vector of plot types:
"traces" or "marginals": traces and 1d marginals of Markov chains of constant parameters, of time-dependent parameters at certain time points (see argument chains.at), chains of log posterior and log observational likelihood values. For selected outputs only, specify traces.constpar, traces.timedeppar, traces.logposterior.
"summary": print summary of acceptance rates and maximum log posterior and log likelihood values.
"pairs": scatterplot matrix of posterior sample of constant parameters.
"time-series": uncertainty range and median time series of time-dependent parameters.
"accept": time series of apparent acceptance frequencies (at the level of thinning).
"realizations": realizations of time-dependent parameters to check for burnin..
"diagnostics": plot diagnostics for inference of time-dependent parameters (note that the plot file could become very large to follow the inference steps).

chains.at

vector of time points at which chains and marginals of time-dependent parameters should be plotted if "traces" or "marginals" is contained in the vector argument type (default: none [numeric(0)]).

labels

optional named vector of expressions to label variables in the plots (names of the expression have to correspond to the variable names as used by the program, expressions can have special symbols, e.g. expression(a=alpha,b=beta,c1=gamma[1])).

units

optional named vector of expressions to add units to variables in the plots (names of the expression have to correspond to the variable names as used by the program, expressions can have special symbols, e.g. expression(a=m^3/s,b=h^-1,c1=m)).

prob.band

probability defining the width of the uncertainty bands plotted for output variables (default value: 0.9)

max.diag.plots

maximum number of diagnostic plots of inference steps

xlim.ts

optional range of time values for time-series plot

n.burnin

number of Markov chain points to omit for density and pairs plots (number of omitted points is max(control$n.adapt,n.burnin)).

nrow

number of plot rows per page (except for pairs plot).

nrow.constpar

number of plot rows per page for traces and marginals (default is nrow).

nrow.timedeppar

number of plot rows per page for time-dependent parameters (default is nrow).

nrow.diagnostics

number of plot rows per page for diagnostics plots (default is nrow).

...

additional arguments passed to the plotting function.