NormalMixture regression

Alternatively, we can also say that the first process distribution is SkewNormal distributed and can use pm.Mixture to define the mixture itself.