產生k個d維的常態分布樣本，產生某個事前機率為P(wi)的常態分布的discriminant

function，計算任兩點間的Euclidean distance及Mahalanobis distance

-Generated k-d-dimensional normal distribution of samples to generate a prior probability P (wi) of the normal distribution of the discriminant function, calculated between any two points in Euclidean distance and Mahalanobis distance