anova.lmcal, anova.nlscal {quantchem}R Documentation

ANOVA tests for calibration models

Description

This function performs ANOVA tests between fitted calibration models.

Usage

anova.lmcal(object, ...)
anova.nlscal(object, ...)

Arguments

object an object of class 'lmcal' or 'nlscal'
... additional arguments (ignored)

Details

For 'lmcal' models this function performs four tests: two ANOVAs between linear and quadratic model, without and with log-log transform (first is mentioned in literature as Mandel's fitting test), and two ANOVAs comparing four polynomial models, weighted and unweighted.

For 'nlscal' models two ANOVAs are performed - between asymptotic models without and with intercept term, and between three and four parameter logistic models.

Value

A list of 'anova' objects:

mandel The test between linear and quadratic fit
logmandel The same test, between two log-log models
table The test between linear, quadratic, cubical and 4th order models without weighting
wtable The test between linear, quadratic, cubical and 4th order models with weighting

Note

The result is printed in readable form, and returned via invisible().

Author(s)

Lukasz Komsta

See Also

lmcal, nlscal

Examples

data(nitrate)
attach(nitrate)
fit = lmcal(conc,area)
anova(fit)
fit2 = nlscal(conc,area)
anova(fit2)

[Package quantchem version 0.12-1 Index]