PREDNASAJUCI / LECTURER :

Daria Dobrycheva

(1) Main Astronomical Observatory of National Academy of
Sciences of Ukraine, Kyiv, Ukraine

NAZOV / TITLE :

Searching for Signs of Exocomets in the TESS Data Using Random Forest
Algorithm

ABSTRAKT / ABSTRACT :

In this study, we demonstrate the effectiveness of the Random Forest
algorithm, a machine learning technique, for identifying potential
exocomet transits within the Sector 1 data of the Transiting Exoplanet
Survey Satellite (TESS). We developed a unique training sample by
incorporating simulated asymmetric transit profiles into observed
light curves, thereby creating realistic data for the model training.
To analyze these light curves, we employed the TSFresh software, which
was a tool for extracting key features that were then used to refine
our Random Forest model training.
Our approach achieved an impressive accuracy of about 96%,
underscoring its capability in separating 'exocomet candidate' and
'non-candidate' in TESS's Sector 1 light curves. This high level of
accuracy, coupled with precision and recall rates exceeding 95% and a
balanced F1-score about 96%, highlights the method's robustness and
reliability. These promising results from Sector 1 motivate us to
extend our analysis across all TESS sectors for detection and study
the comet-like activity in the extrasolar planetary systems.