Validation design and experimental provenance shape apparent machine-learning transferability in legacy Mössbauer spectroscopy
- 1Opticianry Program, Department of Medical Services and Techniques, Vocational School of Health Services, Karamanoğlu Mehmetbey University Karaman Türkiye
- 2Electricity Program, Department of Electricity and Energy, Vocational School of Technical Sciences, Karamanoğlu Mehmetbey University Karaman Türkiye
- 3Medical Physics, University of Wisconsin–Madison Madison WI USA
Abstract
Machine-learning models in chemistry are increasingly trained on values compiled from many published papers and are usually validated by splitting the records at random or by publication; the latter assumes that each publication is an independent experiment. We tested this assumption in a legacy 57Fe Mössbauer corpus. All 45 source publications were re-examined against their original tables and figures, and papers reporting the same alloy composition with the same fitted spectral areas were identified and checked against the sources. Eleven publications reported measurements on three shared specimen series, so the corpus formed 37 experimental-provenance groups containing 171 model-eligible sample states. The target was an operational high-field ferromagnetic spectral fraction, not a martensite phase fraction. At a matched test-set size, median R2 fell from 0.670 under random state-level splitting to 0.322 when publications were held out and 0.287 when provenance groups were held out. With the test states held fixed, retaining related publications in training was associated with a median reduction in apparent error of 6.72 percentage points, which varied strongly between clusters and partly reflects the larger training set. The flexible chemistry-and-temperature model reached a pooled R2 of 0.422 against 0.354 for a calibrated martensite-start comparator but had the lower error in only 17 of 37 groups (group-balanced difference in mean absolute error −1.73 percentage points; 95% CI −7.87 to +4.52). Transferability improved as the evidence base expanded, although provenance-group count could not be isolated as the cause, and a retrospective open-access audit added only 12 model-eligible states. Thus, a publication identifier, which records the paper a value came from, does not guarantee experimental independence, validation design can change the apparent R2 by more than a factor of two, and a pooled advantage need not hold when each experiment is weighted equally. We provide practical recommendations for machine learning on literature-derived chemistry data and release the curated corpus, provenance assignments and code.
Related
- Laser Mössbauer spectroscopy of 229 Th in CaF 2
- Influence of pump-probe beam offset and modulation frequency on photothermal deflection spectroscopy signals in liquid sensing: A combined theoretical and experimental study
- An inadequate over-parameterized Mössbauer spectral analysis for the mistaken detection of surface FeIV=O, in Nature Communications 16, 7642 (2025)
- Experimental Ni II Oscillator Strengths for Transitions from 3d84d to 3d84p Levels Measured Using High-resolution Fourier Transform Spectroscopy
- Disorder-tuned soliton nucleation for vibrational energy harvesting in discrete nonlinear lattices. II. Proof of concept and experimental validation with a piezomagnetic metastructure
BibTeX
@article{armagan2026validation,
title = {{Validation design and experimental provenance shape apparent machine-learning transferability in legacy Mössbauer spectroscopy}},
author = {Osman Armağan and Muhammed Emin Bedi̇r},
journal = {RSC advances},
year = {2026},
month = oct,
doi = {10.1039/d6ra08367f},
url = {https://europepmc.org/article/PMC/PMC13657331},
note = {CC BY-NC},
}
RIS
TY - JOUR
TI - Validation design and experimental provenance shape apparent machine-learning transferability in legacy Mössbauer spectroscopy
AU - Osman Armağan
AU - Muhammed Emin Bedi̇r
PY - 2026
DA - 2026/10/12
JO - RSC advances
DO - 10.1039/d6ra08367f
UR - https://europepmc.org/article/PMC/PMC13657331
AB - Machine-learning models in chemistry are increasingly trained on values compiled from many published papers and are usually validated by splitting the records at random or by publication; the latter assumes that each publication is an independent experiment. We tested this assumption in a legacy 57Fe Mössbauer corpus. All 45 source publications were re-examined against their original tables and figures, and papers reporting the same alloy composition with the same fitted spectral areas were identified and checked against the sources. Eleven publications reported measurements on three shared specimen series, so the corpus formed 37 experimental-provenance groups containing 171 model-eligible sample states. The target was an operational high-field ferromagnetic spectral fraction, not a martensite phase fraction. At a matched test-set size, median R2 fell from 0.670 under random state-level splitting to 0.322 when publications were held out and 0.287 when provenance groups were held out. With the test states held fixed, retaining related publications in training was associated with a median reduction in apparent error of 6.72 percentage points, which varied strongly between clusters and partly reflects the larger training set. The flexible chemistry-and-temperature model reached a pooled R2 of 0.422 against 0.354 for a calibrated martensite-start comparator but had the lower error in only 17 of 37 groups (group-balanced difference in mean absolute error −1.73 percentage points; 95% CI −7.87 to +4.52). Transferability improved as the evidence base expanded, although provenance-group count could not be isolated as the cause, and a retrospective open-access audit added only 12 model-eligible states. Thus, a publication identifier, which records the paper a value came from, does not guarantee experimental independence, validation design can change the apparent R2 by more than a factor of two, and a pooled advantage need not hold when each experiment is weighted equally. We provide practical recommendations for machine learning on literature-derived chemistry data and release the curated corpus, provenance assignments and code.
SN - 2046-2069
ER -