ldt.estim ObjectThis function creates diagnostic plots for estimated regression models of ldt.estim class.
# S3 method for ldt.estim
plot(
x,
equation = 1,
type = c(1, 2, 3, 4, 5, 6),
ablineArgs = list(col = "lightblue"),
textArgs = list(pos = 3, cex = 0.7, col = "red"),
...
)This function creates diagnostic plots for regression models.
It also returns a list with x and y data used in plot functions.
An object of type ldt.estim.
A number or a name of endogenous variable specifying an equation in the estimated system.
One of these numbers: 1, 2, 3, or 5. See which argument in plot.lm documentation.
A list of additional arguments to customize the "text" function used for labeling influential observations.
A list of additional arguments to customize the "abline" function.
additional arguments to be passed to "plot" (or "qqnorm" function for type=2, or "barplot" for type=4).
This function is designed to be similar to plot.lm function.
However, note that an ldt.estim object might be a system estimation.
Some plots use standardized residuals. Note that they are not calculated in a system estimation context. See residuals.ldt.estim documentation for a description. Cook's distance is also calculated equation-wise. Its formula is: $$ d = \frac{r_i^2}{k*var(r)}\frac{h_{ii}}{(1-h_{ii})^2} $$ where \(r_i\) and \(h_{ii}\) are residual and leverage in \(i\)-th observation, respectively. \(var(r)\) is variance of residuals and \(k\) is the number of estimated coefficients in the equation. Note that Cook's distance is not implemented for weighted observations.