def __init__(self, N, B, mu=0.0, sigma=1.0, mu_self=None, sigma_self=None, **kwargs): super(_FixedWeightsMixin, self).__init__(N, B) self._mu = expand_scalar(mu, (N, N, B)) self._sigma = expand_cov(mu, (N, N, B, B)) if (mu_self is not None) and (sigma_self is not None): self._mu[np.arange(N), np.arange(N), :] = expand_scalar(mu_self, (N, B)) self._sigma[np.arange(N), np.arange(N), :] = expand_cov(sigma_self, (N, B, B))
def __init__(self, N, B, mu_0=0.0, sigma_0=1.0, kappa_0=1.0, nu_0=3.0, is_diagonal_weight_special=True, **kwargs): super(_IndependentGaussianMixin, self).__init__(N, B) mu_0 = expand_scalar(mu_0, (B,)) sigma_0 = expand_cov(sigma_0, (B,B)) self._gaussian = Gaussian(mu_0=mu_0, sigma_0=sigma_0, kappa_0=kappa_0, nu_0=max(nu_0, B+2.)) self.is_diagonal_weight_special = is_diagonal_weight_special if is_diagonal_weight_special: self._self_gaussian = \ Gaussian(mu_0=mu_0, sigma_0=sigma_0, kappa_0=kappa_0, nu_0=nu_0)
def __init__(self, T, N, B, mu=0.0, sigma=0.001, mu_self=None, sigma_self=None, **kwargs): super(_FixedWeightsMixin, self).__init__(T, N, B) self._mu = expand_scalar(mu, (N, T, N, B)) self._sigma = expand_cov(sigma, (N, T, N, B, B)) # initialize sigma self._sigma[0, :, 0, :, :] = 0.001 self._sigma[1, :, 1, :, :] = 0.001 self._sigma[1, :, 0, :, :] = 0.001 self._sigma[0, :, 1, :, :] = 0.001 if (mu_self is not None) and (sigma_self is not None): self._mu[np.arange(N), :, np.arange(N), :] = expand_scalar(mu_self, (N, T, B)) self._sigma[np.arange(N), :, np.arange(N), :] = expand_cov(sigma_self, (N, T, B, B))
dim = 2 N = 10 r = 2 + 2 * (np.arange(N) // (N / 2.)) th = np.linspace(0, 4 * np.pi, N, endpoint=False) x = r * np.cos(th) y = r * np.sin(th) L = np.hstack((x[:, None], y[:, None])) # Weight model W = np.zeros((N, N)) # Distance matrix D = ((L[:, None, :] - L[None, :, :])**2).sum(2) sig = np.exp(-D / 2) Sig = np.tile(sig[:, :, None, None], (1, 1, 1, 1)) Mu = expand_scalar(0, (N, N, 1)) for n in range(N): for m in range(N): W[n, m] = npr.multivariate_normal(Mu[n, m], Sig[n, m]) # Adjacency model fig, axs = plt.subplots(4, 5) P = [0.1, 0.3, 0.5, 0.7] i0 = 3 import pickle sm = pickle.load( open('/Users/pillowlab/Dropbox/pyglm-master/Practices/model2.pkl', 'rb')) A = 1 * (np.random.rand(N, N) < P[i0])
def rho(self, value): self._rho = expand_scalar(value, (self.N,))
def mu_b(self, value): self._mu_b = expand_scalar(value, (1,))
def mu_w(self, value): N, B = self.N, self.B self._mu_w = expand_scalar(value, (N, B))
def rho(self, value): self._rho = expand_scalar(value, (self.N, ))
def mu_b(self, value): self._mu_b = expand_scalar(value, (1, ))
def mu_w(self, value): N, B = self.N, self.B self._mu_w = expand_scalar(value, (N, B))
def __init__(self, N, B, rho=0.5, rho_self=None, **kwargs): super(_FixedAdjacencyMixin, self).__init__(N, B) self._rho = expand_scalar(rho, (N, N)) if rho_self is not None: self._rho[np.diag_indices(N)] = rho_self