Plot a fitted MBMM gate curve against one continuous covariate

For a binary (K = 2) gate fit with a single continuous covariate (plus optional intercept), draws the fitted gate probability curve \(P(Z = \code{classes[1]} \mid x)\) over the observed range of that covariate, with subjects' posterior class probabilities overlaid as points (jittered on the y-axis at their MAP class for readability). This is the plot behind Fig. 4/Fig. 6-style panels in Do, McCarthy & Rogers.

Usage

# S3 method for class 'mbmm_gate'
plot(x, focal = NULL, hold_at = list(), n_grid = 200, ...)

Arguments

x

an "mbmm_gate" object from mbmm_fit_gate

focal

name of the continuous covariate to plot on the x-axis. If NULL (default), the first non-intercept term in the gate formula is used.

hold_at

named list of fixed values for any other covariates in the gate formula (only needed if there is more than one covariate)

n_grid

number of points used to draw the fitted curve

...

additional arguments passed to the underlying plot() call (e.g. main, col)

Value

the fitted gate object, invisibly