plot_trend: Plot estimated trend and credible intervals
In jrfaulkner/bnps: Shrinkage-Prior Markov Random Field Smoothing

Description Usage Arguments Details Value See Also

Plot posterior median of trend parameters and associated credible intervals.

plot_trend(theta = NULL, mfit = NULL, obstype = "normal",
  alpha = 0.05, obsvar = NULL, uxvar = NULL, xvar = NULL,
  colset = "blue", trend.col = NULL, trend.lwd = 3, bci.col = NULL,
  pt.col = "gray60", pt.pch = 1, pt.cex = 1, ylab, xlab, main, ylim,
  xlim, ...)

`theta`	A list generated by `extract_theta` function. If `theta` is specified, do not specify a value for `mfit`.
`mfit`	An object containing posterior draws from a `spmrf` or `stan` model fit. The object can be of class `stanfit`, `array`, `matrix`, or `data.frame`. Do not specify `theta` if `mfit` is specified.
`obstype`	Name of probability distribution of observations. This controls the back-transformation of the process parameters. Possible obstypes are 'normal', 'poisson', 'binomial', or 'coalescent'. Assumed link functions are the log for 'poisson' and 'coalescent', and logit for 'binomial'.
`alpha`	Controls level for 100*(1-`alpha`)% Bayesian credible intervals.
`obsvar`	Vector of observations. Assumed to be ordered by time with one observation per time point. If `obsvar` is NULL, then data are not plotted. Note if plotting observations you must also specify a matching `xvar` of the same length.
`uxvar`	vector of unique values of the grid locations of the trend parameters. This can be a numeric or integer vector. Must be the same length as `theta` and must be sorted to match order of locations of `theta` elements. If not specified and no `xvar` is specified, then a sequence from 1 to the length of `theta` with unit spacing will be used. If `xvar` and `obsvar` are specified, then `uxvar` does not need to be specified.
`xvar`	x-variable for plotting observations `obsvar`. Should be a numeric integer vector. Must be the same length as observation vector (`obsvar`). Must be specified if `obsvar` is specified.
`colset`	Color set for trend and credible intervals. These are presets chosen with common color combinations. Choices are 'blue', 'purple', 'red', 'green', and 'gray'. This parameter is overridden If `trend.col` and `bci.col` are specified.
`trend.col`	Color of the trend line. Any standard `R` plotting colors are allowed. Use this along with `bci.col` if color choices other than the presets provided by colset are desired.
`bci.col`	Color of the credible interval band. Any standard `R` colors are allowed.
`pt.col`	Color of data points if plotted with `obsvec` specified.
`pt.pch`	Type of points to use if data plotted with `obsvec` specified (see `points` for details.
`pt.cex`	Size of points to use if data plotted with `obsvec` specified (see `points` for details.
`ylab`	Character string describing label for y-axis. Default is "y" if unspecified.
`xlab`	Character string describing label for x-axis. Default is "x" if unspecified.
`main`	Character string describing label for main title. Default is "" empty string.
`ylim`	Numeric vector with two elements describing limits of y-variable. Default is `range` of combined vector of `obsvec` and extreme values of BCI's if unspecified.
`xlim`	Numeric vector with two elements describing limits of x-variable. Default is `range` of sequence from 1 to max length of elements in `theta` or number of rows in `mfit`.
`...`	Additional parameters passed to the `plot` function.

The input can be either in the form of a list generated by the extract_theta function (passed in as the theta variable) or as a raw model fit object from spmrf or rstan (passed in as mfit). Note that using mfit is more computationally costly than using theta because the extract_theta function is called when mfit is specified.

Returns a plot object.

plot, points, extract_theta, spmrf, stanfit, array, matrix, data.frame