Optimal Transport Based Testing in Factorial Design
Cost Matrix of -Form.
Test multiple linear relationships between probability vectors in fact...
FDOTT: Optimal Transport Based Testing in Factorial Design
Test linear relationships between probability vectors in factorial des...
Check metric properties of cost matrices
Test equality of probability vectors
Compute optimal transport barycenters
Compute optimal transport costs for signed measures
Control FDOTT linear programming solver
Simulations for ot_barycenter_test
Simulations for FDOTT
Generate tabulated samples from probability vectors
Perform optimal transport based tests in factorial designs as introduced in Groppe et al. (2025) <doi:10.48550/arXiv.2509.13970> via the FDOTT() function. These tests are inspired by ANOVA and its nonparametric counterparts. They allow for testing linear relationships in factorial designs between finitely supported probability measures on a metric space. Such relationships include equality of all measures (no treatment effect), interaction effects between a number of factors, as well as main and simple factor effects.