Abstract:In the MiniBooNE technical note TN-112 [1], we reported our preliminary results on νe and π separation by using boosting algorithms. Based on our initial studies, we concluded both boosting algorithms, AdaBoost [2] and -Boost [3], improved the νe and π 0 separation significantly compared with the neural network method. In this note, we use the boosting algorithm to obtain the intrinsic νe CCQE and background separation. Totally 500K cocktail Monte Carlo events and 200K intrinsic νe CCQE events, produced by Steve Brice, were fed into the ’RoeFitters’ reconstruction package[4]. Among them, 88233 intrinsic νe CCQE and 162657 background events passed reconstruction and pre-selection cuts. The pre-selection cuts are listed in the following,

Version 2 Particle Identification in the MiniBooNE Experiment by using the Boosting Algorithm