Sample within rotated feasible parameter space

Using given feasible parameter set, rotate using eigenvectors and sample within rotated hypercube

rotatedSampling(X, samples, expand = 0, ...)

Arguments

X: matrix or data.frame of feasible parameters
samples: Number of samples to take
expand: Ratio with which to expand bounds in rotated parameter space, as a buffer around the original hypercube
...: Arguments to pass to parameterSets

Value

data.frame of parameter values drawn from rotated hypercube

Details

Parameters may not be in feasible parameter set

References

Karel, K. (1990). Membership-set estimation using random scanning and principal component analysis. Mathematics and Computers in Simulation 32(5-6): 535-543 DOI: http://dx.doi.org/10.1016/0378-4754(90)90009-8.

Also possibly related to Iman, R.L. and W.J. Conover (1982). A distribution-free approach to inducing rank correlation among input variables. Communications in Statistics - Simulation and Computation 11(3): 311-334 DOI: http://dx.doi.org/10.1080/03610918208812265.

Author

Joseph Guillaume

Examples



X <- matrix(c(1, 4, 2, 5, 1, 2, 2, 3), ncol = 2)
Y <- rotatedSampling(X, 1e3)
plot(X[, 1], X[, 2], col = "red", cex = 2)
points(Y[, 1], Y[, 2], pch = ".")