This project uses transit data from TESS, the NASA telescope that searches for planets around other stars. When a planet passes in front of its star, the star gets a little dimmer. An eclipsing binary, where two stars pass in front of each other, can make a dip that looks very similar. I am building a classifier that sorts these transit signals into real exoplanets and eclipsing binaries.
- Classifier
- A machine learning model that takes a TESS transit signal and labels it as either a real exoplanet or an eclipsing binary.
- Accuracy study
- Testing how often the classifier gets the label right, so I know how much to trust its results.
- False positive checks
- Checks for signals that the classifier calls a planet when they are not actually a planet.
- Web app
- A Streamlit web app is planned as the final milestone.