Classification¶
The first traditional machine learning method that we will introduce is classification. This is where we try to arrange our data and therefore develop some classification for it. Once these classifications are determined, we can then use them to predict labels for new data that may be produced. For classification to work, our training data must be labelled, we can then test new data coming into our system.
For this, we will look first at decision trees before going on to investigate random forests. You will then put this into practice for the classification of space groups from different materials based on structural and electronic parameters.