Machine learning classification: case of Higgs boson CP state in H to tau tau$ decay at LHC

Machine learning (ML) techniques are rapidly finding a place among the methods of high-energy physics data analysis. Different approaches are explored concerning how much effort should be put into building high-level variables based on physics insight into the problem, and when it is enough to rely on low-level ones, allowing ML methods to find patterns without an explicit physics model. In this paper we continue the discussion of previous publications on the CP state of the Higgs boson measurement of the H→ττ decay channel with the consecutive τ±→ρ±ν; ρ±→π±π0 and τ±→a1±ν; a1±→ρ0π±→3π± cascade decays. The discrimination of the Higgs boson CP state is studied as a binary classification problem between CP even (scalar) and CP odd (pseudoscalar) states using a deep neural network (DNN). Improvements on the classification from the constraints on directly nonmeasurable outgoing neutrinos are discussed. We find that, once added, they enhance the sensitivity sizably, even if only imperfect information is provided. In addition to DNNs we also evaluate and compare other ML methods: boosted trees, random forests, and support vector machines.