Design of Experiments in Python

Description

The py-design package defines the Python module design which implements several routines for the design of experiments. Basically, it serves as a wrapper the Fortran 90 codes for experimental design written by John Burkardt. I have collected, probably, all of them here.

Related Packages

To the best of my knowledge, there is also another Python package implementing several designs called PyDOE. I concentrate more on what is known as randomized designs used in sampling models in order to create surrogate surfaces as well as performing Monte Carlo tasks.

Demos

Here are some demos demonstrating how to use the package:

• demos/demo1.py: Centered Latin Square Design.
• demos/demo2.py: Latin Edge Square Design.
• demos/demo3.py: Latin Random Square Design.
• demos/demo4.py: Adjust a D dimensional dataset of N points so that it forms a Latin hypercube.
• demos/demo5.py: Sparse Grid: Clenshaw Curtis Closed Fully Nested rule.
• demos/demo6.py: Sparse Grid: Fejer 1 Open Fully Nested rule.
• demos/demo7.py: Sparse Grid: Fejer 2 Open Fully Nested rule.
• demos/demo8.py: Sparse Grid: Gauss Patterson Open Fully Nested rule.
• demos/demo9.py: Sparse Grid: Gauss Legendre Open Weakly Nested rule.
• demos/demo10.py: Sparse Grid: Gauss Hermite Open Weakly Nested rule.
• demos/demo11.py: Sparse Grid: Gauss Laguerre Open Non Nested rule.
• demos/demo12.py: Generate the Faure quasirandom sequence.
• demos/demo13.py: Generate the Halton quasirandom sequence.
• demos/demo14.py: Generate the Hammersley quasirandom sequence.
• demos/demo15.py: Generate the Sobol quasirandom sequence.
• demos/demo16.py: Generate the Lambert quasirandom sequence.
• demos/demo17.py: Generate the Improved Distributed Hypercube Sequence.

TODO

• Add references to each algorithm.