SHOW:
|
|
- or go back to the newest paste.
| 1 | basic_model = pm.Model() | |
| 2 | with basic_model: | |
| 3 | ||
| 4 | # Priors for unknown model parameters | |
| 5 | a = pm.Normal('a', mu=10, sd=10, shape=4) | |
| 6 | b = pm.Normal('b', mu=10, sd=10, shape=4) | |
| 7 | c = pm.Normal('c', mu=0, sd=10) | |
| 8 | sigma = pm.HalfNormal('sigma', sd=1) | |
| 9 | ||
| 10 | alpha = np.tile(a, (4, 1)).T | |
| 11 | beta = np.tile(b, (4, 1)) | |
| 12 | offset = np.tile(c, (4, 4)) | |
| 13 | ||
| 14 | # Expected values of outcome | |
| 15 | mu = sigmoid(alpha, beta, offset) # 4x4 matrix here | |
| 16 | ||
| 17 | # Likelihood (sampling distribution) of observations | |
| 18 | # (data also 4x4 matrix) | |
| 19 | Y_obs = pm.Normal('Y_obs', mu=mu, sd=sigma, observed=data) # ERROR! | |
| 20 | ||
| 21 | with basic_model: | |
| 22 | # draw 500 posterior samples | |
| 23 | trace = pm.sample(500) | |
| 24 | ||
| 25 | alpha = pm.Normal('alpha', mu=0, sd=10) | |
| 26 | beta = pm.Normal('beta', mu=0, sd=10, shape=2) | |
| 27 | sigma = pm.HalfNormal('sigma', sd=1) | |
| 28 | mu = alpha + beta[0] * X1 + beta[1] * X2 # mu = Elemwise{add,no_inplace}.0 |
