Neural Network based Jet Charge Tagger
Text for the blessed web page - CDF note 7482

The CDF Collaboration

Date: March 24, 2005

Abstract:

We describe a new version of Jet Charge algorithm for B flavor tagging. We introduce Neural Networks for track and jet probability. We develop a new selection algorithm based on jet probability. Furthermore we make use of the innermost layer of the silicon detector. We measure $\epsilon D^2 = 0.917\pm0.031\%$ on the $e$+SVT sample and $\epsilon D^2 = 0.938\pm0.029\%$ on the $\mu$+SVT sample of 355 pb$^{-1}$.