Using the two scatter matrices approach (SICS) for sliced inversion regression (SIR), the function tests
if the last p-k components have zero eigenvalues, where p is the number of explaining variables. Hence the assumption is that the first k
components are relevant for modelling the response y and the remaining components are not.
SIRasymp(X, y, k, h = 10, ...)A list of class ictest inheriting from class htest containing:
the value of the test statistic.
the p-value of the test.
the degrees of freedom of the test.
character string which test was performed.
character string giving the name of the data.
character string specifying the alternative hypothesis.
the number of non-zero eigenvalues used in the testing problem.
the transformation matrix to the underlying components.
data matrix with the centered underlying components.
the underlying eigenvalues.
the location of the data which was substracted before calculating the components.
a numeric data matrix of explaining variables.
a numeric vector specifying the response.
the number of relevant components under the null hypothesis.
the number of slices used in SIR. Passed on to function covSIR.
other arguments passed on to covSIR.
Klaus Nordhausen
Under the null the first k eigenvalues contained in D are non-zero and the remaining p-k are zero.
For a sample of size \(n\), the test statistic \(T\) is then n times the sum of these last p-k eigenvalue and has under the null a chisquare distribution with \((p-k)(h-k-1)\) degrees of freedom, where h is the number of slices actually used (see covSIR),
therefore it is required that \(k < h-1\).
Nordhausen, K., Oja, H. and Tyler, D.E. (2022), Asymptotic and Bootstrap Tests for Subspace Dimension, Journal of Multivariate Analysis, 188, 104830. <doi:10.1016/j.jmva.2021.104830>.
covSIR, SIRboot
X <- matrix(rnorm(1000), ncol = 5)
eps <- rnorm(200, sd = 0.1)
y <- 2 + 0.5 * X[, 1] + 2 * X[, 3] + eps
SIRasymp(X, y, k = 0)
SIRasymp(X, y, k = 1)
Run the code above in your browser using DataLab