def from_posterior(self, param, samples):
smin, smax = np.min(samples), np.max(samples)
width = smax - smin
x = np.linspace(smin, smax, 100)
y = stats.gaussian_kde(samples.T)(x)
# what was never sampled should have a small probability but not 0,
# so we'll extend the domain and use linear approximation of density on it
x = np.concatenate([[x[0] - 3 * width], x, [x[-1] + 3 * width]])
y = np.concatenate([[0], y, [0]])
return Interpolated(param, x, y)
def from_epdf(param, ambient_counts):
from scipy.stats import gaussian_kde
smin, smax = np.min(ambient_counts), np.max(ambient_counts)
width = smax - smin
x = np.linspace(smin, smax, 200)
y = gaussian_kde(ambient_counts)(x)
x = np.concatenate([[-1e9,-100,0], x, [x[-1] + np.log10(10)]])
y = np.concatenate([[1e-1000,1e-200,1e-2], y, [1e-20]])
sns.scatterplot(x,y)
plt.show()
#return pm.Normal('x', mu=np.mean(ambient_counts), sigma=np.std(ambient_counts))
return Interpolated(param, x, y)