predict.FuNopaRe: Nonparametric Regression for Functional Data

In sipemu/Nonparametric-Functional-Data-Analysis: Nonparametric Functional Data Analysis

Description

predict.FuNopaRe is a function for predictions from the result of nonparametric modell fitting.

Usage

## S3 method for class 'FuNopaRe'
predict(object, newdata, method.params, Bootstrapping = FALSE, ...)

Arguments

object	Afitted object of class inheriting from FuNopaRe
newdata	Matrix with the functional data (curves) each row one for prediction
method.params	Parameters for bootstrapping
Bootstrapping	Using bootstrapping for local adaptive bandwidth selection
...	further arguments passed to or from other methods

Value

FuNopaRe returns an object of the class FuNopaRe; additional with the predictions in the vector Prediction.

Author(s)

Simon Mueller [e-mail]

References

Ferraty, F. and Vieu, P. Nonparametric Functional Data Analysis. Springer 2006.

Rachdi, M. and Vieu, P. Nonparametric regression for functional data: automatic smoothing parameter selection. Journal of Statistical Planning and Inference 137, 9 (2007), 2784-2801.

Benhenni, K., Ferraty, F., Rachdi, M., and Vieu, P. Local smoothing regression with functional data. Computational Statistics 22, 3 (2007) 353???369.

See Also

Semimetric

Examples

# functional data sets
library (fds)

# fat spectrum dataset
Y <- Fatvalues
X <- t(Fatspectrum$y)

# setup semimetric parameters
semimetric.params <- c()
semimetric.params$q <- 2
semimetric.params$nknot  <- 20
semimetric.params$range.grid  <- c (min (Fatspectrum$x), 
                                    max (Fatspectrum$x))

# learn and testsample
learn <- 1:160
test <- 161:215

# parameter estimation and prediction by cross-validation
Learn.Fat.CV <- FuNopaRe (X[learn, ], 
                          Y[learn], 
                          semimetric = "Deriv", 
                          semimetric.params, 
                          bandwidth = "CV")
                          
Predict.Fat.CV <- predict (Learn.Fat.CV, 
                           X[test, ], 
                           method.params = NULL)

plot (Predict.Fat.CV$Prediction, Y[161:215])                          

sipemu/Nonparametric-Functional-Data-Analysis documentation built on May 29, 2019, 10:10 p.m.