def entropy(self, n, p):
x = np.arange(1, np.max(n) + 1)
term1 = n * np.sum(entr(p), axis=-1) - gammaln(n + 1)
n = n[..., np.newaxis]
new_axes_needed = max(p.ndim, n.ndim) - x.ndim + 1
x.shape += (1, ) * new_axes_needed
term2 = np.sum(binom.pmf(x, n, p) * gammaln(x + 1),
axis=(-1, -1 - new_axes_needed))
return term1 + term2
def entropy(self, n, p):
n, p, npcond = self._process_parameters(n, p)
x = np.arange(1, np.max(n) + 1)
term1 = n * np.sum(entr(p), axis=-1) - gammaln(n + 1)
n = n[..., np.newaxis]
new_axes_needed = max(p.ndim, n.ndim) - x.ndim + 1
x.shape += (1, ) * new_axes_needed
term2 = np.sum(binom.pmf(x, n, p) * gammaln(x + 1),
axis=(-1, -1 - new_axes_needed))
return self._checkresult(term1 + term2, npcond, np.nan)
def _entropy(self, p):
return entr(p) + entr(1 - p)
def _entropy(self, n, p):
k = np.arange(n + 1)
vals = self._pmf(k, n, p)
return np.sum(entr(vals), axis=0)
def _entropy(self, p):
# TODO: use logits and binary_cross_entropy_with_logits for more stable
if self.is_logits:
p = expit(p)
return entr(p) + entr(1 - p)
def _entropy(self, n, p):
if self.is_logits:
p = expit(p)
k = np.arange(n + 1)
vals = self._pmf(k, n, p)
return np.sum(entr(vals), axis=0)
