def random(self, n: int, seed: int = None):
if np.isnan(self.params):
raise RuntimeError('Clayton copula parameter cannot be nan')
u = random_uniform(n, self.dim, seed)
if self.params - 1 < 1e-7:
return random_uniform(n, self.dim, seed)
if np.isclose(self.params, 1):
return u
alpha = 1 / self.params
fr = skew_stable.rvs(alpha, beta=1, gamma=cospi2(alpha) ** self.params, pm=1, size=n)
return self.psi(-np.log(u) / fr[:, None])
def random(self, n: int, seed: int = None) -> np.ndarray:
theta = self._theta
if np.isnan(theta):
raise RuntimeError('Clayton copula parameter cannot be nan')
if abs(theta) < 6.06e-6: # magic number
return random_uniform(n, self.dim, seed)
r = random_uniform(n, self.dim, seed)
if self.dim == 2:
r[:, 1] = (r[:, 0] ** (-theta) * (r[:, 1] ** (-theta / (theta + 1)) - 1) + 1) ** (-1 / theta)
return r
else:
gam = rng.standard_gamma(1 / theta, n)[:, None]
r = -np.log(r) / gam
return self.psi(r)
def random(self, n: int, seed: int = None) -> np.ndarray:
theta = self._theta
if np.isnan(theta):
raise RuntimeError('Clayton copula parameter cannot be nan')
if abs(theta) < 1e-7:
return random_uniform(n, self.dim, seed)
r = random_uniform(n, self.dim, seed)
if self.dim == 2:
r[:, 1] = (r[:, 0]**(-theta) *
(r[:, 1]**(-theta / (theta + 1)) - 1) + 1)**(-1 / theta)
return r
else:
gam = rng.standard_gamma(1 / theta, n)[:, None]
r = -np.log(r) / gam
return self.psi(r)
def random(self, n: int, seed: int = None):
u = random_uniform(n, self.dim, seed)
if abs(self.params) < 1e-7:
return u
if self.dim == 2:
v = u[:, 1]
a = -abs(self.params)
v = -1 / a * np.log1p(-v * np.expm1(-a) / (np.exp(-a * u[:, 0]) * (v - 1) - v))
u[:, 1] = 1 - v if self.params > 0 else v
return u
# alpha too large
if log1mexp(self.params) == 0:
return np.ones((n, self.dim))
fr = random_log_series_ln1p(-self.params, n)[:, None]
return self.psi(-np.log(u) / fr)
def random(self, n: int, seed: int = None) -> np.ndarray:
return random_uniform(n, self.dim, seed)
def random(self, n: int, seed: int = None) -> np.ndarray:
return random_uniform(n, self.dim, seed)
def random(self, n: int, seed: int = None):
return random_uniform(n, self.dim, seed)
