residuals.bamlss {bamlss} R Documentation

## Compute BAMLSS Residuals

### Description

Function to compute quantile and response residuals.

### Usage

```## S3 method for class 'bamlss'
residuals(object, type = c("quantile", "response"),
nsamps = NULL, ...)

## S3 method for class 'bamlss.residuals'
plot(x, which = c("hist-resid", "qq-resid", "wp"),
spar = TRUE, ...)
```

### Arguments

 `object` An object of class `"bamlss"`. `type` The type of residuals wanted, possible types are `"quantile"` residuals and `"response"` residuals. `nsamps` If the fitted `bamlss` object contains samples of parameters, computing residuals may take quite some time. Therefore, to get a first feeling it can be useful to compute residuals only based on `nsamps` samples, i.e., `nsamps` specifies the number of samples which are extracted on equidistant intervals. `x` Object returned from function `residuals.bamlss()`. `which` Should a histogram with kernel density estimates be plotted, a qq-plot or a worm plot? `spar` Should graphical parameters be set by the plotting function? `...` For function `residuals.bamlss()` arguments passed to possible `\$residuals()` functions that may be part of a `bamlss.family`. For function `plot.bamlss.residuals()` arguments passed to function `hist.default` and `qqnorm.default`.

### Details

Response residuals are the raw residuals, i.e., the response data minus the fitted distributional mean. If the `bamlss.family` object contains a function `\$mu(par, ...)`, then raw residuals are computed with `y - mu(par)` where `par` is the named list of fitted values of distributional parameters. If `\$mu(par, ...)` is missing, then the fitted values of the first distributional parameter are used.

Randomized quantile residuals are based on the cumulative distribution function of the `bamlss.family` object, i.e., the `\$p(y, par, ...)` function.

### Value

A vector of residuals.

### References

Dunn P. K., and Smyth G. K. (1996). Randomized Quantile Residuals. Journal of Computational and Graphical Statistics 5, 236–244.

van Buuren S., and Fredriks M. (2001) Worm Plot: Simple Diagnostic Device for Modelling Growth Reference Curves. Statistics in Medicine, 20, 1259–1277

`bamlss`, `predict.bamlss`, `fitted.bamlss`.

### Examples

```## Not run: ## Generate data.
d <- GAMart()

## Estimate models.
b1 <- bamlss(num ~ s(x1), data = d)
b2 <- bamlss(num ~ s(x1) + s(x2) + s(x3), data = d)

## Extract quantile residuals.
e1 <- residuals(b1, type = "quantile")
e2 <- residuals(b2, type = "quantile")

## Plots.
plot(e1)
plot(e2)

## End(Not run)
```

[Package bamlss version 1.1-4 Index]