lof {quantchem}R Documentation

Lack-of-Fit testing of calibration models

Description

Performs 'a priori' ANOVA "Lack-of-Fit" tests on fitted calibration models.

Usage

lof(obj)

Arguments

obj An object inheriting from 'cal' (fitted by 'lmcal' or 'nlscal')

Details

This function performs lack-of-fit test on regression residuals. This test assumes, that overall residual error should not be significantly larger than error within groups with the same x (replicates). It is called by 'summary' methods, but also can be called directly by user.

Value

A matrix containing sum of squared residuals, sum of pure error, F-statistic and corresponding p-value.

Note

This test is possible to perform only with minimum 2 replicates of each x value.

Author(s)

Lukasz Komsta

References

see lmcal

See Also

lmcal, nlscal

Examples

x = rep(1:10,10)
y = jitter(x)
fit = lmcal(x,y)
lof(fit)

[Package quantchem version 0.12-1 Index]