Error for loops rStan transformed parameters

Modeling
1

Hello,
I’m trying to estimate the growth parameters of a fish from the capture and recapture data, I want to calculate the age by means of a loops but you leave me error in block transformed parameters

cat(file = “modelhier.stan”, "
data {
int<lower=1> N;// Sample size
int<lower=1> n;// size of id for individuals
int<lower=0> ID[n];// id for individuals
vector[N] dt;// time
vector[N] L;// individuals capture length mm
vector[N] L0;// individuals recapture length mm
}
parameters{
real<lower=0> Linf;//infinity length or max length
real<lower=0> k;//growth rate
real<lower=0> sigma[N];//error
}
transformed parameters {
vector[N] mu;//I want to estimate mean, but I need calculate individual age A[j]

    mu[i] = Linf*(1-exp(-k*(A[ID[i]] + dt[i])));

 for(j in 1:n){
    A[j] = log((L[j]-L0[j])/(L[j]-L0[j]*(exp(-k*dt[j]))))*(-1/k);
  }

}
model{
Linf ~ normal(60,20)T[0,];
k ~ normal(0, 0.1)T[0,];
sigma ~ normal(0,1)T[0,];
for( i in 1:N){
L[i] ~ normal(mu[i], sigma);
}
}
")

dato<-read.csv(“marcaje.csv”)
L<-dato$A1
L0<-dato$A0
dt<-dato$dt
N<-length(dato$ID)
n<-length(unique(dato$ID))
ID<-dato$ID

2

I am not sure that I understand what you are trying to do, but it looks as if the expression for mu[i] should either be inside the for loop or better as a one-line expression like

transformed data {
  vector[N] log_LmL0 = log(L - L0);
}
parameters {
  real<lower = 0> Linf;
  real<lower = 0> k;
  real<lower = 0> sigma[N];
}
transformed parameters {
  vector[N] A = log_LmL0 - log(L - L0 .* exp(-k * dt)) - log(k);
  vector[N] mu = Linf * (1 - exp(-k * (A[ID] + dt)));
}

Then, there is another error that

sigma ~ normal(0, 1)[T[0, ];

is not supported for real arrays. If you don’t need a normalized log-density, you can just omit the T[0, ]; in this case because the truncated area is constant. Or if you want to, you can do

sigma ~ normal(0, 1); // or std_normal() in 2.19
target += -N * log(0.5); // due to the truncation
3

Thanks bgoodri, I have changed according to what you recommend, I need to calculate the age of each individual “A [j]” in function of its id therefore I put the for loop in n, since the individuals have been recaptured and measured (but I do not know the parameter "k "), the problem is on the line transformed data {…

cat(file = “modelhier.stan”, "
data {
int<lower=1> N;// Sample size
int<lower=1> n;// size of id for individuals
int<lower=0> ID[n];// id for individuals
vector[N] dt;// time
vector[N] L;// individuals capture length mm
vector[N] L0;// individuals recapture length mm
}

parameters{
real<lower=0> Linf;//infinity length or max length
real<lower=0> k;//growth rate
real<lower=0> sigma[N];//error
//real<lower=0> mu[N];
}
transformed data {
vector[n] A;
for(j in 1:n){    
A[j] = log((L[j]-L0[j])/(L[j]-L0[j]*(exp(-k*dt[j]))))*(-1/k);
}

}
transformed parameters {
vector[N] mu;
for(i in 1:N){
mu[i] = Linf*(1-exp(-k*(A[ID[i]] + dt[i])));
}
}

model{
Linf ~ normal(60,20)T[0,];
k ~ normal(0, 0.1)T[0,];
sigma ~ normal(0,1);

    for( i in 1:N){
    L[i] ~ normal(mu[i], sigma);

}
}
")

4

If a function depends on k and k is declared in the parameters block, then it all needs to go into the transformed parameters block rather than transformed data. I think what I wrote before was equivalent to what you were trying to do.

5

Thanks once again,given the change,

cat(file = “modelhier.stan”, "
data {
int N;
vector[N] dt;
vector[N] L;
vector[N] L0;
}
transformed data {
vector[N] log_LmL0 = log(L - L0);
}
parameters {
real<lower = 0> Linf;
real<lower = 0> k;
real<lower = 0> sigma;//[N];
}
transformed parameters {
vector[N] A = -(log_LmL0 - log(L - L0 .* exp(-k * dt)))/k;
vector[N] mu = Linf * (1 - exp(-k * (A + dt)));
}
model{
Linf ~ normal(60,20)T[0,];
k ~ normal(0, 0.1)T[0,];
sigma ~ normal(0, 1);
//target += -N * log(0.5);
for( i in 1:N){
L[i] ~ normal(mu[i], sigma);
}
}
")

I get
Rejecting initial value:
Error evaluating the log probability at the initial value.

Initialization between (-2, 2) failed after 100 attempts.
[1] “Error in sampler$call_sampler(args_list[[i]]) : Initialization failed.”
[1] “error occurred during calling the sampler; sampling not done”

I suppose that transformed parameters A , gave me log na

6

Its good, the problem was with log() gave nas values, I run for N observations, but not for n individuals that was repet in all meditions (N). I want to iterations for n individuals. Only this. Thanks