Figure 4
Confusion matrix of the (a) CNN ° XGBoost and (b) I ° RF methods on the whole data set. These confusion matrices are normalized by the number of predictions, so that the sum of a row is equal to 1. As an example, on (a) 93% of the prolates were well classified, 4% were confused with oblates, 2% with core shell spheres and 1% with spheres. |