fda.usc (version 0.9.8.1)

predict.fregre.plm: Predict method for semi-functional linear regression model.

Description

Computes predictions for functional (and non functional) explanatory variables and scalar response using asymmetric kernel estimation.

Usage

## S3 method for class 'fregre.plm':
predict(object,newx=NULL,\dots)

Arguments

object

fregre.plm object.

newx

An optional data list in which to look for variables with which to predict. If omitted, the fitted values are used. List of new explanatory data.

...

Further arguments passed to or from other methods.

Value

• Return:
• fitted.values
  Predited scalar response.

Details

Predicts from a fitted fregre.plm object. The first item in the data list of newx argument is called "df" and is a data frame with the response and non functional explanatory variables, as lm. Functional variable (fdata and fd class) is introduced in the second item in the data list of newx argument.

References

Aneiros-Perez G. and Vieu P. (2005). Semi-functional partial linear regression. Statistics {&} Probability Letters, 76:1102-1110. Ferraty, F. and Vieu, P. (2006). Nonparametric functional data analysis. Springer Series in Statistics, New York. Hardle, W. Applied Nonparametric Regression. Cambridge University Press, 1994.

See Also

See Also as: fregre.plm.

Examples

data(tecator)
ind=1:129
x=tecator$absorp.fdata[ind,]
newx=tecator$absorp.fdata[-ind,]
y=tecator$y$Fat[ind]
newy=tecator$y$Fat[-ind]
dataf=as.data.frame(tecator$y[ind,])

f=Fat~Water+x
ldata=list("df"=dataf,"x"=x)
res.plm=fregre.plm(f,ldata)
ndataf=as.data.frame(tecator$y[-ind,])
nldata=list("df"=ndataf,"x"=newx)
pred.plm=predict.fregre.plm(res.plm,nldata)

## with 2nd derivative of functional data
# x.fd=fdata.deriv(x,nderiv=2)
# f2=Fat~Protein+x.fd
# ldata2=list("df"=dataf,"x.fd"=x.fd)
# res.plm2=fregre.plm(f2,ldata2)
# newx.fd=fdata.deriv(newx,nderiv=2)
# nldata2=list("df"=ndataf,"x.fd"=newx.fd)
# pred.plm2=predict.fregre.plm(res.plm2,nldata2)

# MEP.plm=((1/length(newy))*sum((newy-pred.plm)^2))/var(newy)
# MEP.plm2=((1/length(newy))*sum((newy-pred.plm2)^2))/var(newy)
# par(mfrow=c(1,2))
# boxplot(newy-pred.plm,ylim=c(-5,5))
# boxplot(newy-pred.plm2,ylim=c(-5,5))