def CreateAsPrior(cls, argDict, Data):
''' Creates Gaussian-Wishart prior for params that generate Data.
Returns GaussWishDistr object with same dimension as Data.
Provided argDict specifies prior's expected covariance matrix
and expected mean
'''
D = Data.dim
m = np.zeros(D)
dF = np.maximum( argDict['dF'], D+2)
kappa = argDict['kappa']
ECovMat = WishartDistr.createECovMatFromUserInput(argDict, Data)
invW = ECovMat * (dF - D - 1)
return cls(dF=dF, kappa=kappa, m=m, invW=invW)
def CreateAsPrior(cls, argDict, Data):
""" Creates Gaussian-Wishart prior for params that generate Data.
Returns GaussWishDistr object with same dimension as Data.
Provided argDict specifies prior's expected covariance matrix
and expected mean
"""
D = Data.dim
m = np.zeros(D)
dF = np.maximum(argDict["dF"], D + 2)
kappa = argDict["kappa"]
ECovMat = WishartDistr.createECovMatFromUserInput(argDict, Data)
invW = ECovMat * (dF - D - 1)
return cls(dF=dF, kappa=kappa, m=m, invW=invW)
def logWishNormConst(self): return WishartDistr.calc_log_norm_const(self.logdetW(), self.dF, self.D)
def logWishNormConst(self):
return WishartDistr.calc_log_norm_const(self.logdetW(), self.dF, self.D)