This repository contains analysis for the topic "Machine Learning for the Planetary Albedo".
Out of the several models tested, XGBoostRegressor gives the least Root Mean Squared Error (RMSE) ~ 0.03.
Lunar - Actual Map
The XGBoostRegressor gives RMSE value ~ 0.16. Other regressors, for example the LinearSVR, give similar results as the XGBoostRegressor. For the case of mercury, the chemical composition maps were predicted before predicting the albedo maps.
Mercury - Actual Map
Lunar Albedo Map Prediction
Mercury Element Map Prediction
Mercury Albedo Map Prediction
