plsRglm (version 0.7.6)

loglikpls: loglikelihood function for plsR models

Description

This function provides loglikelihood computation for an univariate plsR model.

Usage

loglikpls(residpls, weights=rep.int(1,length(residpls)))

Arguments

residpls

Residuals of a fitted univariate plsR model

weights

Weights of observations

Value

• real
  Loglikelihood value

Details

Loglikelihood functions for plsR models with univariate response.

References

Baibing Li, Julian Morris, Elaine B. Martin, Model selection for partial least squares regression, Chemometrics and Intelligent Laboratory Systems 64 (2002) 79-89. http://dx.doi.org/10.1016/S0169-7439(02)00051-5

See Also

AICpls for AIC computation and logLik for loglikelihood computations for linear models

Examples

data(pine)
ypine <- pine[,11]
Xpine <- pine[,1:10]
(Pinscaled <- as.data.frame(cbind(scale(log(ypine)),scale(as.matrix(Xpine)))))
colnames(Pinscaled)[1] <- "yy"

lm(yy~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,data=Pinscaled)

modpls <- plsR(log(ypine),Xpine,10)
modpls$Std.Coeffs
lm(yy~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,data=Pinscaled)

AIC(lm(yy~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,data=Pinscaled))
print(logLik(lm(yy~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,data=Pinscaled)))

sum(dnorm(modpls$RepY, modpls$Std.ValsPredictY, sqrt(mean(modpls$residY^2)), log=TRUE))
sum(dnorm(Pinscaled$yy,fitted(lm(yy~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,data=Pinscaled)),sqrt(mean(residuals(lm(yy~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,data=Pinscaled))^2)), log=TRUE))
loglikpls(modpls$residY)
loglikpls(residuals(lm(yy~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,data=Pinscaled)))
AICpls(10,residuals(lm(yy~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,data=Pinscaled)))
AICpls(10,modpls$residY)