Search for Standard Model $H \rightarrow \tau \tau$ Decays using a new $\tau$ identification algorithm

Clean and efficient identification of $\tau$ leptons is required for a number CDF Run~II analyses such as $W \rightarrow \tau \nu$ and $Z \rightarrow \tau \tau$ production cross section measurements, top quark dilepton studies, and searches for both SUSY particles and MSSM Higgs bosons. Improved $\tau$ lepton identificantion can also be used to help improve the sensitivity of standard model Higgs boson searches by increasing the number of final states available to the analyses. We present a new $\tau$ lepton identification method based on a Boosted Decision Tree (BDT) approach, which like other multivariate methods can provide higher selection efficiency as a function of the associated fake rate with respect to what one obtains using a rectangular set of selection criteria. The performance of the new algorithm, which gives an improvement on order of 15\% over the current CDF $\tau$ identification algorithm, will be shown.