A neural net based algorithm to identify b and c jet with the CDF detector

A sophisticated vertexing algorithm is used to select all the displaced vertices and charged particles produced in the decays of heavy flavor hadrons present in b and c quark jets. The information on tracks and vertices is then used in a neural net to enhance the b to light-jet and b to c-jet discrimination. An additional neural net uses information from traditional tagging algorithms used in CDF (including soft leptons) to produce a final discriminant variable for b and c jets identification. The procedure to validate the algorithm using CDF data will be discussed in this talk.