Figure 5
Feature selection using recursive cABC analysis in sensory and genomic data for pain. The dataset includes subject gender, pain thresholds to heat, cold, blunt pressure, punctate pressure (von Frey hairs), and electrical stimuli with and without prior sensitization by local application of capsaicin or menthol cream, and genetic information on 29 common variants in eight human genes reported to modulate pain, including single nucleotide variants and haplotypes, acquired from n = 125 healthy young volunteers48. (A) Variable importance according to a 5 × 20 nested cross-validation feature selection using random forests and the generic permutation importance provided in the "permutation_importance" method of the "sklearn.inspection" package. Bar colors indicate the selection of informative variables in different repeated selection steps using cABC analysis, from light blue = "not selected" to dark blue and black for features selected in deeper until the last repetition of cABC analysis. (B) and (C) Results of the cABC analysis of the mean variable importance. The ABC plots (blue lines) show the cumulative distribution function of the importance variables together with the identity distribution, xi = constant (magenta line), and the uniform distribution, i.e., as a stopping criterion for the repetitions of the cABC analysis. The red lines show the boundaries between the ABC subsets "A", "B" and "C". The figure was created using Python version 3.8.13 for Linux (https://www.python.org) with the seaborn statistical data visualization package (https://seaborn.pydata.org22) and our Python package "cABCanalysis" (https://pypi.org/project/cABCanalysis/).
