Invertible Diagnostics: A Topological and Spectral Framework for Multi-Variable Attribution
Standard clinical triage relies on binary classification trees, which frequently misclassify overlapping conditions by forcing continuous phenomena into rigid, mutually exclusive categories. Invertible Diagnostics resolves this by replacing flat binary boundaries with a colour-mapped Möbius manifold and sliding spectral calibration. In short, simply by overlapping colours, we may alter the saturation of each colour, defining the contribution any given variable makes to the overall condition. We map out this basic geometric modelling, demonstrating its highly practical application through an objective clinical analysis of orthostatic hypotension—separating primary post-sepsis vascular dysfunction from contributory diabetic autonomic neuropathy
Authors
- HN Brookhouse
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-12
- DOI
- https://doi.org/10.5281/zenodo.24159897
- Primary Topic
- Artificial Intelligence in Healthcare
- Type
- preprint