predict.pls {PLSiMCpp}

Predict according to the Estimated Parameters

Description

Predict Y based on new observations.

Usage

## S3 method for class 'pls'
predict(object, x_test = NULL, z_test, ...)

Arguments

object

fitted partially linear single-index model, which could be obtained by

x_test

input matrix (linear covariates of test set).

z_test

input matrix (nonlinear covariates of test set).

...

additional arguments.

plsim.MAVE, or plsim.est, or plsim.vs.soft.

Value

y_hat

prediction.

Examples

n = 50
sigma = 0.1

alpha = matrix(1, 2, 1)
alpha = alpha/norm(alpha, "2")

beta = matrix(4, 1, 1)

x = matrix(1, n, 1)
x_test = matrix(1,n,1)

z = matrix(runif(n*2), n, 2)
z_test = matrix(runif(n*2), n, 2)

y = 4*((z%*%alpha-1/sqrt(2))^2) + x%*%beta + sigma*matrix(rnorm(n),n,1)
y_test = 4*((z_test%*%alpha-1/sqrt(2))^2) + x_test%*%beta + sigma*matrix(rnorm(n),n,1)


# Obtain parameters in PLSiM using Profile Least Squares Estimator
fit_plsimest = plsim.est(x, z, y)

preds_plsimest = predict(fit_plsimest, x_test, z_test)

# Print the MSE of the Profile Least Squares Estimator method
print( sum( (preds_plsimest-y_test)^2)/nrow(y_test) )

# Obtain parameters in PLSiM using Penalized Profile Least Squares Estimator
fit_plsim = plsim.vs.soft(x, z, y,lambda = 0.01)

preds_plsim = predict(fit_plsim, x_test, z_test)

# Print the MSE of the Penalized Profile Least Squares Estimator method
print( sum( (preds_plsim-y_test)^2)/nrow(y_test) )