Exoplanet Transit Detection
Faculty supervisor: Kevin Johnston
I develop a Python pipeline to preprocess and label stellar light-curve data, train a one-dimensional convolutional neural network on confirmed positive and negative observations, and apply the model to uncertain candidates.
- 01 / PREPARELight-curve data
Preprocess and label stellar observations.
- 02 / CLASSIFYA one-dimensional CNN
Train, evaluate, and apply the model to candidates.
- 03 / UNDERSTANDReview the errors
Trace false positives and negatives by TIC ID.
Light-curve visualizations support manual review and model refinement, keeping individual observations visible behind the classification results.
