cvreg: Cross Validation and Robust Estimation Utilities

Description Usage Arguments Value Examples

Plot marginal effects / partial dependence curves

1	margeffPlot(fit, X, y = NULL, predfun, wch = 1, res = 100, points = FALSE)

`fit`	a model it
`X`	a data frame to be used. should not included the response variable.
`y`	the response variable. used only for plotting numeric covariates, and is optional.
`predfun`	a prediction function that takes the arguments fit and newdata, and returns a vector of predicted values.
`wch`	which variables to plot. can be a single number indexing the variable, or a vector of two numbers to plot the predicted values as a function of both variables.
`res`	the resolution of the plot, in other words, how many points along the function should define the curve/line.
`points`	if TRUE, and y is not NULL, the data points are plotted with the regression line.

a plot

1
2
3

fit <- KIR(prog ~ .-sex, diabetes)
yhat <- predict(fit, newdata, type = "response")
margeffPlot(fit = fit, X = diabetes[,-c(2, 11)], predfun = yhat, wch = 4)