A unified method for computing probabilities of misclassification when the covariance matricesare known
Atsu S.S. Dorvlo · Communications in Statistics - Simulation and Computation · 1993
The Qos and Qm are two leading The Linear and Quadratic discriminant functions are shown to be expressible as linear combinations of independent non-central chi-square random variables when the covariance matrices are known. Hence the probabilityof misclassification associated with them could be computed using numerical integration procedures.