plot.deepgof1: Plot the residual map of a DeepGOF-1 test

In ebrahim.gof: Goodness-of-Fit and Calibration Tests for Logistic Regression

Plot the residual map of a DeepGOF-1 test

Description

Draws the K by K residual map that the network scored in deepgof1 or deepgof1.external, with the value of every cell written in it. A cell holds the observed minus the expected number of events among the records whose two covariates fall in that pair of rank groups, divided by its null standard deviation, so under a correct model the cells are roughly standard normal. Vermilion cells had more events than the model predicted, blue cells fewer.

Usage

## S3 method for class 'deepgof1'
plot(x, colour = TRUE, digits = 1, ...)

Arguments

x	A deepgof1 object.
colour	FALSE shades the cells in greys by the size of the value only; the sign is then read from the number in the cell.
digits	Decimals shown in the cells. Default 1.
...	Not used.

Details

The axes are the rank groups of the two covariates the map was drawn over (the axes element of the result), 1 being the lowest. With one covariate the map is K * K quantile cells along its ranks, read row by row from the top left. The p-value is that of the test; the map says where the misfit sits, and a single large cell is not a test.

Value

x, invisibly.

See Also

deepgof1, deepgof1.external.

Examples

set.seed(1)
n  <- 400
x1 <- rnorm(n); x2 <- rnorm(n)
y  <- rbinom(n, 1, plogis(0.5 * x1 + 0.5 * x2 + 0.8 * x1 * x2))
fit <- glm(y ~ x1 + x2, family = binomial())
plot(deepgof1(fit, B = 49))

ebrahim.gof documentation built on Oct. 11, 2026, 5:07 p.m.