FriesischScott / UncertaintyQuantification.jl

Uncertainty Quantification in Julia
MIT License
28 stars 7 forks source link

Design of Experiments #86

Closed FriesischScott closed 1 year ago

FriesischScott commented 1 year ago

Now that the response surface was implemented in #87 we should also add the most used design of experiments such as:

ExperimentalDesign.jl could be a solution but it has a lot of dependencies and currently set the compat for Julia to 1.8.

FriesischScott commented 1 year ago

Since the designs are defined on [-1, 1] the generated samples have to be mapped into the physical space. The general approach would be to use cdf(Uniform(-1, 1), x to go to cdf space, then with quantile(Normal(), x) to go to standard normal space. From the to_physical_space! should do the trick.