Neural network analysis for gamma gamma -> 3pi at Daphne

Abstract

We consider the possibility of using neural networks in experimental data analysis in Daphne. We analyze the process $\gamma\gamma\to \pi^+ \pi^- \pi^0$ and its backgrounds using neural networks and we compare their performances with traditional methods of applying cuts on several kinematical variables. We find that the neural networks are more efficient and can be of great help for processes with small number of produced events.