How to model time-dependent variables in pymc3

Questions
5

Thank you for your answers!

I was wondering if you know of an easy way to do incorporate a more complex function, specifically a skewed normal distribution.

I am trying to fit the function that describes a skewed normal distribution to some data because this function appears to fit well. While there is a SkewNormal in pymc3, it is used as a probability function and I am not sure how to use pymc3 distribution as a function.

Normally, in Python I would define the function as

def skew(x,e=0,w=1,a=0, mag=1):
     t = (x-e) / w
    return 2 * mag * scipy.stats.norm.pdf(t) * scipy.stats.norm.cdf(a*t)

However, this includes a definite integral when not using scipy function.

If you know a good way to:
a) implement this function
b) reuse the scipy functions within pymc3
c) reuse the pymc3 function

I would be very grateful.