
My comment would be: try also model P1 and P2 as a normal distribution, and do tt.exp(M_det1) and tt.exp(M_det2) for the mu of Poission. This is usually a more standard way (i.e., Generalized linear model)

My comment would be: try also model P1 and P2 as a normal distribution, and do tt.exp(M_det1) and tt.exp(M_det2) for the mu of Poission. This is usually a more standard way (i.e., Generalized linear model)