LINEAR AND NON LINEAR DISCRIMINANT FUNCTIONS FOR THE CLASSIFICATION OF HIGH ENERGY PHYSICS DATA

To classify high energy physics events, we propose to use linear and non linear discriminant functions. Global event shape and connected new morphological variables are considered. Three kinds of non linear discriminant functions are designed. All these have proven to be more efficient classifier than the linear functions. The efficiencies and purities achieved with the non linear classifiers are in average 1 to 7.