betanalpha
commented
7 years ago Good catch — a fix has been pushed. Thanks!
On Jan 23, 2017, at 4:40 PM, Evan Eskew notifications@github.com wrote:
Awesome resource Michael. However, I think there's a conflict between your description of the non-centered implementation of the Eight Schools model and the actual Stan code. The code states that "theta[j] = mu + tau * theta_tilde[j]" whereas the written description of the model has sigma in place of tau. Figured it would be worth the clarification for those (like myself) who are just getting familiar with coding non-centered versions of models.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/betanalpha/knitr_case_studies/issues/1, or mute the thread https://github.com/notifications/unsubscribe-auth/ABdNliX7j9SHc_U7fMcVvjI_mpOdx4W3ks5rVR5fgaJpZM4LrlEn.
Awesome resource Michael. However, I think there's a conflict between your description of the non-centered implementation of the Eight Schools model and the actual Stan code. The code states that "theta[j] = mu + tau * theta_tilde[j]" whereas the written description of the model has sigma in place of tau. Figured it would be worth the clarification for those (like myself) who are just getting familiar with coding non-centered versions of models.