ExperimentalDesign provides tools for Design of Experiments in Julia, enabling the construction of designs for screening, modeling, exploration, and optimization.
Development on this package is ongoing, so expect things to change. Pull requests are more than welcome!
Check the documentation for the latest features and API, and check the examples directory for Jupyter Notebooks and code.
Current features are:
- Designs that support categorical and continuous factors
- Integration with StatsModels
- Full factorial designs:
- Explicit: for small designs that fit in memory
- Iterable: for larger designs, generates experiments on demand
- Two-level fractional factorial designs
- Plackett-Burman designs for screening (check the example)
- Box-Behnken and central composite designs for response surface modeling
- Flexible random designs using the Distributions package
- Latin Hypercube designs using the LatinHypercubeSampling.jl package
- Several variance-optimizing criteria