Plots the "deviance residuals" from a logit.spike model.

1 2 | ```
PlotLogitSpikeResiduals(model, ...)
PlotProbitSpikeResiduals(model, ...)
``` |

`model` |
A model object inheriting from |

`...` |
Additional arguments to be passed to |

The "deviance residuals" are defined as the signed square root each observation's contribution to log likelihood. The sign of the residual is positive if half or more of the trials associated with an observation are successes. The sign is negative otherwise.

The "contribution to log likelihood" is taken to be the posterior mean of an observations log likelihood contribution, averaged over the life of the MCMC chain.

The deviance residual is plotted against the fitted value, again averaged over the life of the MCMC chain.

The plot also shows the .95 and .99 bounds from the square root of a chi-square(1) random variable. As a rough approximation, about 5% and 1% of the data should lie outside these bounds.

Steven L. Scott

1 2 3 4 5 6 7 8 9 10 11 12 13 14 | ```
simulate.logit.spike <- function(n = 100, p = 10, ngood = 3,
niter=1000){
x <- cbind(1, matrix(rnorm(n * (p-1)), nrow=n))
beta <- c(rnorm(ngood), rep(0, p - ngood))
prob <- plogis(x %*% beta)
y <- runif(n) < prob
x <- x[,-1]
draws <- logit.spike(y ~ x, niter=niter)
plot.ts(draws$beta)
return(invisible(draws))
}
model <- simulate.logit.spike()
plot(model, "fit")
plot(model, "fit", scale = "probability", number.of.buckets = 15)
``` |

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.