testLoess.Rd
Calculates an F test to evaluate significant differences between a LOESS
model and a parametric alternative estimated with lm
testLoess(lmobj, loessobj, alpha = 0.05)
lmobj | An object of class |
---|---|
loessobj | An object of class |
alpha | Desired Type I error rate of test. |
Printed output describing the results of the test.
data(Prestige, package="carData") linmod <- lm(prestige ~ income, data=Prestige) lomod <- loess(prestige ~ income, data=Prestige) testLoess(linmod, lomod)#> F = 4.92 #> Pr( > F) = 0 #> LOESS preferred to alternative