Modular classification systems for antiparticle studies in PAMELA

In this study, a modular classification system is proposed in order to improve the signal/background discrimination of the imaging calorimeter of PAMELA experiment. For the pre-processing phase, we have used the recently developed Super Paramagnetic Clustering algorithms, while for the classification phase, we have applied supervised Artificial Neural Networks and Support Vector Machines. Results obtained using simulated data show performances of the modular systems adequate for the experimental task of the detector