FunChisq: Chi-Square and Exact Tests for Non-Parametric Functional Dependencies
Statistical hypothesis testing methods for non-parametric functional dependencies using asymptotic chi-square or exact statistics. These tests were motivated to reveal evidence for causality based on functional dependencies. They include asymptotic functional chi-square tests, an exact functional test, a comparative functional chi-square test, and also a comparative chi-square test. The normalized non-constant functional chi-square test was used by Best Performer NMSUSongLab in HPN-DREAM (DREAM8) Breast Cancer Network Inference Challenges. For continuous data, these tests offer an advantage over regression analysis when a parametric functional form cannot be assumed; for categorical data, they provide a novel means to assess directional dependencies impossible with classical Pearson's chi-square or Fisher's exact tests.