MOP: Mean Orthogonal Projection (Combine Dimension Reducton...

In abnormally-distributed/cvreg: Cross Validation and Robust Estimation Utilities

Description

Mean Orthogonal Projection (Combine Dimension Reducton Subspaces)

Usage

MOP(p, method = c("const", "sqrtinv"))

Arguments

p	a list of symmetric projection matrices.
method	whether constant weights or sqrt(1/rank) weights should be used.

Value

a list

References

Liski E., Nordhausen K., Oja H., and Ruiz-Gazen A. (2016), Combining Linear Dimension Reduction Subspaces. In: Recent Advances in Robust Statistics: Theory and Applications. doi: 10.1007/9788132236436_7.

Liski, E., Nordhausen, K., Oja, H., & Ruiz-Gazen, A. (2012). Averaging orthogonal projectors. https://arxiv.org/pdf/1210.2575.pdf

Examples

fit1 = SIR(bmi ~ ., diabetes[,-2], rank = 1)
fit2 = SIR(bmi ~ ., diabetes[,-2], rank = 3)
fit3 = SIR(bmi ~ ., diabetes[,-2], rank = 5)
fit4 = SIR(bmi ~ ., diabetes[,-2], rank = 9)
MOP(p = list(fit1$sirmat, fit2$sirmat, fit3$sirmat, fit4$sirmat), method = "sqrtinv")

abnormally-distributed/cvreg documentation built on May 3, 2020, 3:45 p.m.