Multidimensional Likelihood-based Flavor Tagging at D{\O}

We study the determination of the initial state flavor of $b$ quarks produced in $p\bar{p}$ collisions at $\sqrt{s}=1.96$ TeV recorded by the D{\O} detector operating at the Fermilab Tevatron. This is vital to the observation of $B_{s}$ mixing. A multidimensional likelihood method has been developed using kinematic properties of particles that originate from the other $b$ quark produced in events with reconstructed $B$ mesons. The performance of the tagging algorithm is determined by measuring the $B_d$ mixing frequency with a data sample of semileptonic $B$ meson decays from which dilutions and efficiencies are extracted. The method can be easily extended through the addition of other kinematic variables.