Manuscript Error: Possible needless slicing in example?

[paw-lu]

Description

To be honest I'm a little unsure here, but in 5.10

with pm.Model() as sp_individual:
    τ = pm.HalfCauchy("τ", 1)
    β = pm.Normal("β", mu=0, sigma=τ, shape=(Bi.shape[1], groups))
    μi = pm.Deterministic("μi", pm.math.dot(Bi, β))
    σ = pm.HalfNormal("σ", 1)
    up = pm.Normal("up", μi[:,index].T.ravel(), σ, observed=uptake)
    idata_sp_individual = pm.sample(2000, return_inferencedata=True)

where

index = range(12)
groups = len(index)

I'm a bit confused on this line:

    up = pm.Normal("up", μi[:,index].T.ravel(), σ, observed=uptake)

specifically the μi[:, index] statement. If groups = len(index), μi has len(index) columns. So if we slice it using μi[:, index], then I think we do nothing to it? we're essentially saying: Take all rows (:) and all columns (index since μi has len(index) rows) of μi. So μi[:, index] == μi, and the slicing does nothing.

Reference

5.7 Listings 5.10

Author Section

Do not close until

• Added to Errata
• Fixed in Open Access
• Fixed in latex source