Optimized search for single-top-quark production at the Fermilab Tevatron

We use a neural-network technique to search for standard model single-top-quark production in the $106{\mathrm{pb}}^{-1}$ dataset accumulated by the Collider Detector at Fermilab detector during the 1992–1995 collider run (“run I”). Using a sample of 64 $W+1,$ 2, 3 jets events, we set a 95% confidence level upper limit of 24 pb on the W-gluon and $W^{*}$ combined single-top cross section.