CODEWITHSUNDEEP
rsyntax-errorbayesianmodel-fitting
Bakaburg
Bakaburg

Reputation: 3311

Error when fitting Bayesian model with the Rethinking R package

I'm trying to fit a very simple model to estimate the prevalence of disease as described here https://www.ecdc.europa.eu/sites/default/files/documents/Methodology-estimating-point-prevalence%20-SARS-CoV-2-infection-pooled-RT-PCR-testing.pdf on page 6, using the Rethinking R package

Here's my code:

quap(alist(
    p ~ dbeta(.3, .3),
    p_test ~ 1 - dbinom(0, s, p), # I tried also p_test <- 1 - dbinom(0, s, p)
    k ~ dbinom(w, p_test)
), data = list(s = 10, k = 30, w = 200))

but I receive the error: Error in pars[[i]] : subscript out of bounds in quap

What am I doing wrong?

Upvotes: 1

Views: 195

Answers (1)

Lstat
Lstat

Reputation: 1460

Reordering of definitions in alist does the trick.

quap(
 alist(
    k ~ dbinom(w, p_test),
    p ~ dbeta(.3, .3),
    p_test ~ 1 - dbinom(0, s, p) 
 ), 
 data = list(s = 10, k = 30, w = 200)
)

This code returns:

Quadratic approximate posterior distribution

Formula:
k ~ dbinom(w, p_test)
p ~ dbeta(0.3, 0.3)
p_test ~ 1 - dbinom(0, s, p)

Posterior means:
         p 
0.01575975 

Log-likelihood: -2.55

Upvotes: 2

Related Questions