A multivariate approach to heavy flavour tagging with cascade training

This paper compares the performance of artificial neural networks and boosted decision trees, with and without cascade training, for tagging b-jets in a collider experiment. It is shown, using a Monte Carlo simulation of WH → lνqq̄ events, that boosted decision trees outperform artificial neural networks. Furthermore, cascade training can substantially improve the performance of both boosted decision trees and artificial neural networks. The results show that, for a b-tagging efficiency of 60%, the light jet rejection power given by boosted decision trees is about 25% higher than that given by artificial neural networks. The cascade training technique improves the performance of artificial neural networks by about 30% and the performance of boosted decision trees by about 20%.