RG-NFT: analysis pipeline for "How Far Ahead Is Driver Drowsiness Predictable?

Analysis pipeline accompanying the manuscript "Participant-independent
characterisation of the prediction horizon for driver drowsiness in two
physiological corpora".

Contains feature extraction for electrooculography, electrocardiography and
electroencephalography; a nested leave-one-subject-out evaluation harness;
participant-level bootstrap and paired inference; the reliability-gated
neuro-fuzzy fusion model (RG-NFT) and its ablations; and the numbered scripts
that reproduce every table and figure in the paper. Machine-readable result
files are included so that any reported value can be traced to its source.

No participant data are redistributed. Both corpora are obtained separately
from their own repositories: DD-Database (doi:10.5061/dryad.5tb2rbp9c, CC0)
and Cao et al. 2019 (doi:10.6084/m9.figshare.6427334, CC BY 4.0).