A simple example for continuous inputs
Assume that the input vectors for each class are from a
Gaussian distribution, and all classes have the same
covariance matrix.
For two classes, C1 and C0, the posterior is a logistic: