In the library the parent distribution mean value for the betas is a single number, but we can use a number of features for our model. Some of those features might have a positive effect and some might be negative, and I am wondering if it is correct to group them all under one mean value
Basically what you have written is:
mu_beta = pm.Normal('mu_beta', mu=0, sd=100)sigma_beta = pm.HalfNormal('sigma_beta', sd=100)
In the library the parent distribution mean value for the betas is a single number, but we can use a number of features for our model. Some of those features might have a positive effect and some might be negative, and I am wondering if it is correct to group them all under one mean value
Basically what you have written is:
mu_beta = pm.Normal('mu_beta', mu=0, sd=100)
sigma_beta = pm.HalfNormal('sigma_beta', sd=100)
And my thoughts are that it should be:
So instead of having a single mu_beta for all of the features, each feature has a unique mu_beta.
Let me know what your thoughts are on this.