Data analysis of matched and unmatched hittracks performed now
Trained RNN chooses 94.8% of tracks that actually come back correct (72% correct over total data)
Try of a RNN classifier in Keras to choose which particles actually recurl into detector => No logic found with different architectures
Try of XGBOOST classifier to choose which particles actually recurl into detector => No logic found with different architectures
No correlation to be found about which particles of the proposed ones actually recurl