Unknown variable: vector error message

Modeling
1

I am completely new to Stan, so sorry for my standard questions. I am building a bayesian multi-level time series model and the code below is the start of my project. However, when I want to run it I get the error message: Unknown variable: vector variable “vector” does not exist. This is at line 66 in the model part for vector[H] beta;. I don’t know what I did wrong. Can someone help me with this error? Thanks in advance!

data {
  int<lower=0> N; // Observations
  int<lower=0> Npreds;
  int<lower=0> H; // persons
  int<lower=0> id[N]; // ID variable
  real y[N]; // Y Vector
  matrix [N, Npreds] x;
}

parameters {
  real beta[H];
  matrix [H, Npreds] gamma;
  real gamma1;
  real gamma2;
  real<lower=0> y_sig;   // regression variance
}

model {

  //Priors
  y_sig~uniform(0,1000); 
  vector[H] beta;
    for(h in 1:H)
      beta[h] = gamma1 + gamma2*col(x,1)[h]; 
  
  // Likelihood
  for(n in 1:N) 
  y[n] ~ normal(row(beta, id[n]) * row(x,n)', y_sig);
}

Another question I have is regarding the brms package in R. Does anyone know whether you could also fit models with a panel structure (multiple individuals over time) with it? I can’t find anything about it.

2

Variable / Parameter declarations have to be the first thing in a block. So the line: vector[H] beta; has to occur before y_sig~uniform(0,1000);. Also, you have two things called beta in this model. One in the parameters block and one in the model block.

3 Likes
3

Thanks! this solved my problem!