Machine Learning Validation of TESS Transit Light Curves: Case Study of TOI 7325.01

This paper uses machine learning to sort transit candidates from TESS into likely exoplanets or false positives. We took features such as transit depth, duration, average flux, and flux variation from TESS light curves to train a Random Forest model. The model then determines the probability of how likely a candidate is to be an actual exoplanet, thereby assisting in identifying which targets to observe more attentively. For instance, the model gave TOI-7325.01 a probability of 0.73, suggesting a strong likelihood of being a real exoplanet. The results show that basic light curve features can give useful advice early on when checking if an exoplanet is real. Future research could add in a wider range of stellar attributes along with larger data sets to improve its predictive accuracy and use the method on more TESS Objects of Interest.