| image.bma {BMS} | R Documentation |
Plots a grid with signs and inclusion of coefficents vs. posterior model probabilites for the best models in a 'bma' object:
## S3 method for class 'bma': image(x, yprop2pip = FALSE, order.by.pip = TRUE, do.par = TRUE, do.grid = TRUE, do.axis = TRUE, cex.axis = 1, ...)
x |
a list of class bma (cf. bms for further details) |
yprop2pip |
if yprop2pip=TRUE then the grid lines on the vertical axis are scaled according to the coefficients' inclusion probabilites.If yprop2pip=FALSE (default) then the gridlines on the vertical axis are equidistant.
|
order.by.pip |
with order.by.pip=TRUE (default), coefficents are sorted according to their posterior inclusion probabilites along the vertical axis. If order.by.pip=FALSE they are ordered as they were provided to bms. |
do.par |
Defaults to do.par=TRUE, which adjusts par()$mar for optimal positioning. Set do.par=FALSE for customizing par yourself. |
do.grid |
do.grid=TRUE (default) plots grid lines among the chart's boxes, akin to the low level command grid. do.grid=FALSE omits the gridlines. |
do.axis |
do.axis=TRUE (default) plots axis tick marks and labels (cf. axis). do.axis=FALSE omits them. |
cex.axis |
font size for the axes (cf. axis), defaults to 1 |
... |
Parameters to be passed on to image.default. |
Under default settings, blue corresponds to positive sign, red to a negative sign, white to non-inclusion.
Martin Feldkircher and Stefan Zeugner
coef.bma for the coefficents in matrix form, bms for creating 'bma' objects.
Check http://bms.zeugner.eu for additional help.
data(datafls)
model=bms(datafls,nmodel=200)
#plot all models
image(model,order.by.pip=FALSE)
image(model,order.by.pip=TRUE,cex.axis=.8)
#plot best 7 models, with other colors
image(model[1:7],yprop2pip=TRUE,col=c("black","lightgrey"))