We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

physics.chem-ph

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Physics > Chemical Physics

Title: The long road to calibrated prediction uncertainty in computational chemistry

Authors: Pascal Pernot
Abstract: Uncertainty quantification (UQ) in computational chemistry (CC) is still in its infancy. Very few CC methods are designed to provide a confidence level on their predictions, and most users still rely improperly on the mean absolute error as an accuracy metric. The development of reliable uncertainty quantification methods is essential, notably for computational chemistry to be used confidently in industrial processes. A review of the CC-UQ literature shows that there is no common standard procedure to report nor validate prediction uncertainty. I consider here analysis tools using concepts (calibration and sharpness) developed in meteorology and machine learning for the validation of probabilistic forecasters. These tools are adapted to CC-UQ and applied to datasets of prediction uncertainties provided by composite methods, Bayesian Ensembles methods, machine learning and a posteriori statistical methods.
Subjects: Chemical Physics (physics.chem-ph); Data Analysis, Statistics and Probability (physics.data-an)
Journal reference: J. Chem. Phys. 156, 114109 (2022)
DOI: 10.1063/5.0084302
Cite as: arXiv:2201.01511 [physics.chem-ph]
  (or arXiv:2201.01511v2 [physics.chem-ph] for this version)

Submission history

From: Pascal Pernot [view email]
[v1] Wed, 5 Jan 2022 09:31:28 GMT (9539kb,D)
[v2] Fri, 18 Mar 2022 08:38:54 GMT (3910kb,D)

Link back to: arXiv, form interface, contact.