pymc3 generate stochastic variables with array of

问题:

In pymc3, a stochastic variable of array shape say 3 can be generated as follows

y = Normal('y', mu, sigma, shape=3, observed=some_data)

Now suppose that y depends on an array of parameters mu = [1,2,3] and sigma = [4,5,6] instead of single values, how would I specify it?

回答1:

I wanted a way to do the something like following

for i in range(3):
    y[i] = Normal(mu[i], sigma[i], observed=somedata[i]) 

The way to do it is to simply pass in a list of values.

y=Normal(mu, sigma, observed=somedata) 

where mu and sigma are lists