predict.ldblm {dbstats}

Predicted values for a ldblm object

Description

predict.ldblm returns the predicted values, obtained by evaluating the local distance-based linear model in the new data (newdata2), using newdata1 to estimate the "kernel weights".

Usage

                                  
## S3 method for class 'ldblm'
predict(object,newdata1,newdata2=newdata1,
        new.k.knn=3,type.var="Z",...)

Arguments

object

an object of class ldblm. Result of ldblm.

newdata1

data.frame or matrix which contains the values of Z (if type.var="Z". The squared distances between k new individuals and the original n individuals (only if type.var="D2"). Finally, the G inner products matrix (if type.var="G"). newdata1 is used to compute kernels and local weights.

newdata2

the same logic as newdata1. newdata2 is used to compute the Distance-based Regressions with (dblm). If newdata2=NULL, newdata2 <- newdata1.

new.k.knn

setting a minimum bandwidth in order to check that a candidate bandwidth h doesn't contains DB linear models with only one observation. If new.h.knn=NULL, takes the distance that includes the 3 nearest neighbors for each new individual row.

type.var

set de type of the newdata paramater. Can be "Z" if newdata contains the values of the explanatory variables, "D2" if contains the squared distances matrix or "G" if contains the inner products matrix.

...

arguments passed to or from other methods to the low level.

Value

A list of class predict.ldblm containing the following components:

fit

predicted values for the k new individuals.

newS

matrix (with dimension (k,n)) of weights used to compute the predictions.

Note

Look at which way (or type.var) was made the ldblm call. The parameter type.var must be consistent with the data type that is introduced to ldblm.

Author(s)

Boj, Eva <[e-mail]>, Caballe, Adria <[e-mail]>, Delicado, Pedro <[e-mail]> and Fortiana, Josep <[e-mail]>

References

Boj E, Delicado P, Fortiana J (2010). Distance-based local linear regression for functional predictors. Computational Statistics and Data Analysis 54, 429-437.

Boj E, Grane A, Fortiana J, Claramunt MM (2007). Selection of predictors in distance-based regression. Communications in Statistics B - Simulation and Computation 36, 87-98.

Cuadras CM, Arenas C, Fortiana J (1996). Some computational aspects of a distance-based model for prediction. Communications in Statistics B - Simulation and Computation 25, 593-609.

Cuadras C, Arenas C (1990). A distance-based regression model for prediction with mixed data. Communications in Statistics A - Theory and Methods 19, 2261-2279.

Cuadras CM (1989). Distance analysis in discrimination and classification using both continuous and categorical variables. In: Y. Dodge (ed.), Statistical Data Analysis and Inference. Amsterdam, The Netherlands: North-Holland Publishing Co., pp. 459-473.

See Also

ldblm for local distance-based linear models.

Examples

# example to use of the predict.ldblm function

n <- 100
p <- 1
k <- 5

Z <- matrix(rnorm(n*p),nrow=n)
b1 <- matrix(runif(p)*k,nrow=p)
b2 <- matrix(runif(p)*k,nrow=p)
b3 <- matrix(runif(p)*k,nrow=p)

s <- 1
e <- rnorm(n)*s

y <- Z%*%b1 + Z^2%*%b2 +Z^3%*%b3 + e

D <- as.matrix(dist(Z))
D2 <- D^2

newdata1 <- 0

ldblm1 <- ldblm(y~Z,kind.of.kernel=1,method="GCV",noh=3,k.knn=3)
pr1 <- predict(ldblm1,newdata1)
print(pr1)
plot(Z,y)
points(0,pr1$fit,col=2)
abline(v=0,col=2)
abline(h=pr1$fit,col=2)