def read_data(self, path): conn = sql.connect(path) self.V = pd.read_sql('select * from data;', conn).values self.S = data.euclidean_to_simplex(self.V) self.Yl = data.angular_to_euclidean(data.euclidean_to_angular(self.V)) self.A = data.euclidean_to_angular(self.Yl) self.Vi = self.cast_to_cube(self.A) self.pVi = self.probit(self.Vi) conn.close() return
def read_data(self, path): self.Z = pd.read_csv(path).values self.R = self.Z.max(axis=1) self.V = (self.Z.T / self.R).T self.S = data.euclidean_to_simplex(self.V) self.A = data.euclidean_to_angular(self.V) self.Yl = data.angular_to_euclidean(self.A) self.Vi = self.cast_to_cube(self.A) self.pVi = self.probit(self.Vi) self.I = (np.arange(self.Z.shape[0]), ) return
def generate_posterior_predictive(self, n_per_sample = 10): """ Generates posterior prediction, projects to hypercube, then casts to angular space """ hyp = self.generate_posterior_predictive_hypercube(n_per_sample) return dm.euclidean_to_angular(hyp)
def generate_posterior_predictive_angular(self, n_per_sample = 1): hypercube = self.generate_posterior_predictive_hypercube(n_per_sample) return euclidean_to_angular(hypercube)