Prediction of the Fatigue Strength of Steel Based on Interpretable Machine Learning.
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: GBT predicts steel fatigue strength from 437 NIMS records; 5-fold CV and 80/20 holdout. Composition-only R²=.92/MAPE 8.15%; adding heat treatment yields .98/3.14%; atomic features .98/3.05%. Feature-selection/SHAP/SR identify TT, Mo and related inputs.
Technology: GBT predicts steel fatigue strength from 437 NIMS records; 5-fold CV and 80/20 holdout. Composition-only R²=.92/MAPE 8.15%; adding heat treatment yields .98/3.14%; atomic features .98/3.05%. Feature-selection/SHAP/SR identify TT, Mo and related inputs.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- GBT predicts steel fatigue strength from 437 NIMS records; 5-fold CV and 80/20 holdout. Composition-only R²=.92/MAPE 8.15%; adding heat treatment yields .98/3.14%; atomic features .98/3.05%. Feature-selection/SHAP/SR identify TT, Mo and related inputs.
- Limitation / still unknown
- Retrospective database prediction, not newly run fatigue tests; performance depends on the NIMS sample and split.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://doi.org/10.3390/ma16237354 · https://www.mdpi.com/1996-1944/16/23/7354
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.