Hi, Aki and I were playing around with some simple examples and found things that pedantic mode wasn’t catching. See below.
Andrew
–
First we created a model with three parameters, two of which had no prior:
parameters {
real a;
real b;
real c;
}
model {
a ~ normal(0, 1);
}
Pedantic mode returned: “Warning: The parameter c was declared but was not used in the density calculation.”
But what about the parameter b?
Here’s another:
data {
real y;
}
parameters {
real a;
real b;
real c;
}
model {
y ~ normal(a + b + c, 1);
}
This time, pedantic returned, “Warning: The parameter c has no priors.
Warning: The parameter b has no priors.”
What happened to a?