This repository contains the code and resources for an Exoplanet Detection System using Machine Learning and Deep Learning. The goal of this project is to detect exoplanets, which are planets that orbit stars outside our solar system, using machine learning and deep learning techniques. The system takes in astronomical data, such as light curves or spectra, and uses various machine learning algorithms to classify the data into different categories, such as exoplanet candidates or false positives.
ML models used are Decision Trees, Random Forest and XGB.
A CNN model is used for binary classification.
The dataset has been taken from Kaggle
XGB and Random Forest showed an accuracy of 99.8%, while CNN showed 99.4% and 99.1% for two different test sets and Decision Trees showed 95% accuracy. Future work would include applying LSTM as dataset is a time-series data and working on CNN model.