A multivariate analysis approach for the imaging atmospheric Cherenkov telescopes system H.E.S.S.

A multivariate particle classification approach is proposed and applied to the analysis of the data from H.E.S.S. (High Energy Stereoscopic System). The combination of results from the three shower reconstruction methods: Hillas, model and 3D-model, leads to a substantial gain in the discrimination power between photons and hadrons. The construction and use of a combined effective estimator improves by several factors the signal-to-background ratio which is extremely important in case of studies of the faint and extended sources. The results of this approach are presented for a typical set of sources. The consequent gain in the sensitivity is shown through a comparison to the H.E.S.S. published results.