Machine Learning Model for the Planetary Albedo

The goal of the project is to use ML techniques to identify relationships between planetary mapped datasets, with the goal of providing deeper understanding of planetary surfaces and to have predictive power for planetary surfaces with incomplete datasets. There are three main proposals for the stated problem:

1. Explore various ML models, adjust the parameters and compare their performance. During the preliminary work, I already explored several basic ML models; however, we should try more advanced ones and perform a deeper comparison depending on the parameters and metrics.
2. Predict chemicals from chemicals. Based on preliminary work, we see that although in Mercury the correlation between albedo and chemicals is low, the correlationships between some chemicals are pretty high. Thus, in order to fill the "gaps" in the chemical maps, we could try to predict one chemical based on some others.
3. Perform convolutional prediction. Rather than using only one corresponding pixel for the prediction, we could include the surrounding pixels since they might offer additional information.