def __init__(self, likelihoods_list, gp_link=None ,name='heterogeneous_likelihood'):
if gp_link is None:
gp_link = link_functions.Identity()
super(HetLikelihood, self).__init__(gp_link=gp_link, name=name)
self.likelihoods_list = likelihoods_list
def __init__(self,
Y_metadata,
gp_link=None,
noise_mult=1.,
known_variances=1.,
name='Scaled_het_Gauss'):
if gp_link is None:
gp_link = link_functions.Identity()
if not isinstance(gp_link, link_functions.Identity):
print(
"Warning, Exact inference is not implemeted for non-identity link functions,\
if you are not already, ensure Laplace inference_method is used")
# note the known_variances are fixed, not parameterse
self.known_variances = known_variances
self.noise_mult = Param('noise_mult', noise_mult,
Logexp()) # Logexp ensures its positive
# this is a parameter, so it gets optimized, gradients calculated etc.
#super(ScaledHeteroscedasticGaussian, self).__init__(gp_link, variance=1.0, name=name)
super(Gaussian, self).__init__(gp_link, name=name)
# note: we're inheriting from Likelihood here, not Gaussian, so as to avoid problems with the Gaussian variance.
#add a new parameter by linking it (see just above in GPy.likelihoods.gaussian.Gaussian).
self.link_parameter(self.noise_mult)
if isinstance(gp_link, link_functions.Identity):
self.log_concave = True
def __init__(self, sigma=None, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
if sigma is None:
self.sigma = 0.5
else:
self.sigma = sigma
super(Gaussian, self).__init__(gp_link, name='Gaussian')
def __init__(self,gp_link=None, deg_free=5, sigma2=2):
if gp_link is None:
gp_link = link_functions.Identity()
super(HetStudentT, self).__init__(gp_link, name='Hetro_Student_T')
self.v = Param('deg_free', float(deg_free), Logexp())
self.link_parameter(self.v)
self.v.constrain_fixed()
self.log_concave = False
def __init__(self, Y, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
self.Y = Y
Clases = np.unique(self.Y)
auxC = np.prod(Clases)
if auxC > 0:
K = len(np.unique(self.Y))
else:
K = len(np.unique(self.Y)) - 1
self.K = K
super(MultiClassMAA, self).__init__(gp_link, name='MultiClassMAA')
def __init__(self, K, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
super(Categorical, self).__init__(gp_link, name='Categorical')
self.K = K
def __init__(self, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
super(Beta, self).__init__(gp_link, name='Beta')
def __init__(self, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
super(Poisson, self).__init__(gp_link, name='Poisson')
def __init__(self, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
super(Exponential, self).__init__(gp_link, name='Exponential')
def __init__(self, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
super(HetGaussianMA, self).__init__(gp_link, name='HetGaussianMA')
def __init__(self, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
super(Student, self).__init__(gp_link, name='Student')
def __init__(self, likelihood_fns=[], name='Mixed_noise'):
#TODO Why do we need to specify a link function?
super(Mixed, self).__init__(name=name,
gp_link=link_functions.Identity())
self.likelihood_fns = likelihood_fns
def __init__(self, gp_link=None):
if gp_link is not None:
raise NotImplementedError, "this likelihood assumes a complicated pair of link functions..."
super(HetGP, self).__init__(link_functions.Identity(), 'HetGP')
def __init__(self, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
super(BinaryClassMA, self).__init__(gp_link, name='BinaryClassMA')
def __init__(self, gp_link=None):
if gp_link is None:
gp_link = link_functions.Identity()
super(Dirichlet, self).__init__(gp_link, name='Dirichlet')
def __init__(self,gp_link=None, deg_free=5, sigma2=2):
if gp_link is None:
gp_link = link_functions.Identity()
super(HetBeta, self).__init__(gp_link, name='Hetra_beta')
self.log_concave = False
