Log normal sampling

lognormal.py

import pystan
import numpy
import matplotlib.pyplot as plt

y = numpy.random.lognormal(mean = 0.7, sigma = 1.0, size = 5)

# Both of these models produce the same results
model_code = """
data {
  int<lower=1> N;
  vector<lower=0.0>[N] y;
}

parameters {
  real mu;
  real<lower=0.0> sigma;
}

model {
  y ~ lognormal(mu, sigma);
}
"""

sm1 = pystan.StanModel(model_code = model_code)

model_code = """
data {
  int<lower=1> N;
  vector<lower=0.0>[N] y;
}

parameters {
  real mu;
  real<lower=0.0> sigma;
}

model {
  log(y) ~ normal(mu, sigma);
}
"""

sm2 = pystan.StanModel(model_code = model_code)

# These fits should be equivalent
fit1 = sm1.sampling(data = { 'N' : len(y), 'y' : y }, iter = 100000, warmup = 1000)
fit2 = sm2.sampling(data = { 'N' : len(y), 'y' : y }, iter = 100000, warmup = 1000)

print fit1
print fit2