# Thermal data attribution and reuse

Source: **Kirgsn (wkirgsn), Electric Motor Temperature, version 3, 26 April 2021**. Measurements collected at the Chair of Power Electronics and Electrical Drives (LEA), Paderborn University.

- Dataset: https://www.kaggle.com/datasets/wkirgsn/electric-motor-temperature
- Dataset license: **Creative Commons Attribution–ShareAlike 4.0 International (CC BY-SA 4.0)**.
- License: https://creativecommons.org/licenses/by-sa/4.0/
- Legal terms: https://creativecommons.org/licenses/by-sa/4.0/legalcode
- Original authors’ research context: https://github.com/upb-lea/deep-pmsm

The archived version-3 dataset metadata explicitly records CC BY-SA 4.0. The original authors’ repository licenses its code separately; that does not change the dataset license. The seven fitted candidates in this companion are Dynamical’s separate investigation, not results attributed to that repository’s authors.

## Adaptations in this export

Dynamical fitted and froze seven candidates in the retained investigation. This export reproduces existing predictions and scores; it does not fit or run a new model. Changes to the dataset presentation include selection of the 14 held-out sessions, addition of frozen model predictions and development bands, calculated errors and metrics, time in seconds/minutes, extrema-preserving selection of display samples, and rounding display values to six decimal places. Original sample indices remain in each trace. Figures connect selected original samples; omitted samples are not used to recompute metrics.

The adapted thermal data in `data/` and the corresponding data figures are provided under **CC BY-SA 4.0**. Keep the source and Dynamical attribution, link to the license, identify your changes, and retain ShareAlike terms when distributing adaptations. Do not imply endorsement by the dataset contributors.

This notice is not a blanket license over application code, prose, trademarks, fonts, supplier documents or unrelated studies. Fonts have their own OFL notices in `../assets/fonts/`. No calibrated sensor-uncertainty record accompanies these metrics; the error envelopes are development-derived, not measurement uncertainty.

Suggested figure credit: “Data: Kirgsn (wkirgsn), LEA, Paderborn University. Frozen predictions and adapted display data: Dynamical. CC BY-SA 4.0. Display samples preserve extrema; metrics use original samples.”
