Lightcurve Classification for Periodic Sources using the Catalina Real-Time Transients Survey (CRTS)
@kvmu @rdegardner @rachzili13 @yaoshi1994 @brooke1313
We present a small study in the classification of (periodic) stars from the CRTS dataset. Using a Random Forest classifier we achieved a 81.59% classification accuracy on 16 classes of stars.
Raw lightcurve data was analyzed to extract features using the FATS module (Feature Analysis for Time Series), developed by Isadora Nun (github: @isadoranunand) Pavlos Protopapas from the Institute of Applied Computational Science. This is the FATS paper.
The classes of stars that we considered were, see paper:
- ACEP
- Beta-Lyrae
- Blazhko
- Cep II
- EA
- ELL
- EW
- HADS
- Hump
- LADS
- LPV
- PCEB
- RRab
- RRc
- RRd
- RS CVn