A dataset for training and testing machine learning models to predict Time-Temperature-Transformation (TTT) curves in uranium-molybdenum (U-Mo-X) alloys. The data includes training, test, virtual, and calibration sets, along with elemental descriptors and principal component values for the alloys. It was authored by Sunidhi Garg from the University of Virginia and last updated on April 20, 2026.
Use Cases
- Training physics-regularized ML models to predict TTT curves based on alloy composition data.
- Testing model performance on unseen U-Mo-X alloy compositions using the provided test set.
- Generating predictions for virtual alloy compositions absent from the training data.
- Calibrating model predictions using a dedicated calibration dataset for U-10Mo alloy.
Strengths
- Includes distinct datasets for training, testing, virtual prediction, and calibration, suggesting a structured ML workflow.
- Provides both raw elemental descriptors and derived principal component values for the alloys.
- Authored by a researcher at the University of Virginia and hosted on a Dataverse platform.
Limitations
- Description metadata is limited; actual data quality, column definitions, and row counts require manual inspection after download.
- Column-level documentation is absent; field semantics must be inferred from the file contents.
Provenance
- Source
- University of Virginia (Libra Data) Harvested Dataverse
- Collection Method
- Likely generated from computational materials science research for ML model development.
- Time Range
- null
- Freshness
- Last updated 2026-04 20 14:46:26; freshness should be verified.
- Geography
- null