Exoplantets anommally detection with autoencoders with ~99% accuracy
Code imports a csv from: PHL-EC data set in github This repository contains the PHL-EC dataset used in the classification problem for exoplanets. The data set is used in the manuscript "Habitability Classification of Exoplanets: A Machine Learning Insight"
The aim is to detect which planets are habitable versus the vast majority of planets that are not.
To do so, 3 structures are presented: 1.- PCA + autoencoder. (99% accuracy) 2.- deep autoencoder .(96% accuracy) 3.- autoencoder (dimension reduction) autoencoder (anomaly detection)