Scaling LEO lightcurve classification: hyperparameter-tuned boosted decision trees on an expanded dataset
RAS Techniques and Instruments · 06.11.2026
DOI: 10.1093/rasti/rzag095 ↗Kurz gefasst
Fachartikel in der Zeitschrift RAS Techniques and Instruments, online erschienen am 06.11.2026. Verfasst von Billy Shrive und 9 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Building on previous low Earth orbit (LEO) satellite classification work, this paper aims to investigate the hyperparameter space of MMT-9 lightcurves and apply the optimal hyperparameters to classification. Both inter-platform and intra-platform lightcurve classification tasks are considered. The best hyperparameters were found to consist of 500 boosting steps, a learning rate of 0.01 and a max tree depth of 6 nodes. These hyperparameters were then used for classification with a lightcurve dataset size of 1000 lightcurves per platform, an order of magnitude greater than previous work, resulting in a slight increase in classification accuracy (87.92% versus 85.65%), suggesting a plateau was already being approached with 100 lightcurves per platform.
Zitieren
APA
Shrive, B., Pollacco, D., Chote, P., Blake, J. A., Cooke, B. F., McCormac, J., West, R., Airey, R., MacManus, A., & Whitlock, P. (2026). Scaling LEO lightcurve classification: hyperparameter-tuned boosted decision trees on an expanded dataset. RAS Techniques and Instruments. https://doi.org/10.1093/rasti/rzag095
Vancouver
Shrive B, Pollacco D, Chote P, Blake JA, Cooke BF, McCormac J, et al. Scaling LEO lightcurve classification: hyperparameter-tuned boosted decision trees on an expanded dataset. RAS Techniques and Instruments. 2026. doi:10.1093/rasti/rzag095
BibTeX
@article{Shrive2026,
author = {Shrive, Billy and Pollacco, Don and Chote, Paul and Blake, James A and Cooke, Benjamin F and McCormac, James and West, Richard and Airey, Robert and MacManus, Alex and Whitlock, Phineas},
title = {Scaling LEO lightcurve classification: hyperparameter-tuned boosted decision trees on an expanded dataset},
journal = {RAS Techniques and Instruments},
year = {2026},
doi = {10.1093/rasti/rzag095}
}
EUVIMED