apunkt/GPy
1
0
Fork 0
mirror of https://github.com/SheffieldML/GPy.git synced 2026-05-21 14:05:14 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fixed likelihood tests for new parameters structure

Browse source
This commit is contained in:
Alan Saul 2014-02-07 15:16:52 +00:00
parent c28f11f291
commit 186feb45a1
6 changed files with 207 additions and 168 deletions
Download patch file Download diff file Expand all files Collapse all files

24
GPy/likelihoods/likelihood.py
View file

@ -13,12 +13,12 @@ from ..core.parameterization import Parameterized
class Likelihood(Parameterized):
"""
Likelihood base class, used to defing p(y|f).
Likelihood base class, used to defing p(y|f).
All instances use _inverse_ link functions, which can be swapped out. It is
expected that inherriting classes define a default inverse link function
To use this class, inherrit and define missing functionality.
To use this class, inherrit and define missing functionality.
Inherriting classes *must* implement:
pdf_link : a bound method which turns the output of the link function into the pdf
@ -27,7 +27,7 @@ class Likelihood(Parameterized):
To enable use with EP, inherriting classes *must* define:
TODO: a suitable derivative function for any parameters of the class
It is also desirable to define:
moments_match_ep : a function to compute the EP moments If this isn't defined, the moments will be computed using 1D quadrature.
moments_match_ep : a function to compute the EP moments If this isn't defined, the moments will be computed using 1D quadrature.
To enable use with Laplace approximation, inherriting classes *must* define:
Some derivative functions *AS TODO*
@ -36,7 +36,7 @@ class Likelihood(Parameterized):
"""
def __init__(self, gp_link, name):
super(Likelihood, self).__init__(name)
super(Likelihood, self).__init__(name)
assert isinstance(gp_link,link_functions.GPTransformation), "gp_link is not a valid GPTransformation."
self.gp_link = gp_link
self.log_concave = False
@ -44,6 +44,10 @@ class Likelihood(Parameterized):
def _gradients(self,partial):
return np.zeros(0)
def update_gradients(self, partial):
if self.size > 0:
raise NotImplementedError('Must be implemented for likelihoods with parameters to be optimized')
def _preprocess_values(self,Y):
"""
In case it is needed, this function assess the output values or makes any pertinent transformation on them.
@ -303,7 +307,7 @@ class Likelihood(Parameterized):
"""
TODO: Doc strings
"""
if len(self._get_param_names()) > 0:
if self.size > 0:
link_f = self.gp_link.transf(f)
return self.dlogpdf_link_dtheta(link_f, y, extra_data=extra_data)
else:
@ -314,7 +318,7 @@ class Likelihood(Parameterized):
"""
TODO: Doc strings
"""
if len(self._get_param_names()) > 0:
if self.size > 0:
link_f = self.gp_link.transf(f)
dlink_df = self.gp_link.dtransf_df(f)
dlogpdf_dlink_dtheta = self.dlogpdf_dlink_dtheta(link_f, y, extra_data=extra_data)
@ -327,7 +331,7 @@ class Likelihood(Parameterized):
"""
TODO: Doc strings
"""
if len(self._get_param_names()) > 0:
if self.size > 0:
link_f = self.gp_link.transf(f)
dlink_df = self.gp_link.dtransf_df(f)
d2link_df2 = self.gp_link.d2transf_df2(f)
@ -345,9 +349,9 @@ class Likelihood(Parameterized):
#Parameters are stacked vertically. Must be listed in same order as 'get_param_names'
# ensure we have gradients for every parameter we want to optimize
assert dlogpdf_dtheta.shape[1] == len(self._get_param_names())
assert dlogpdf_df_dtheta.shape[1] == len(self._get_param_names())
assert d2logpdf_df2_dtheta.shape[1] == len(self._get_param_names())
assert dlogpdf_dtheta.shape[1] == self.size
assert dlogpdf_df_dtheta.shape[1] == self.size
assert d2logpdf_df2_dtheta.shape[1] == self.size
return dlogpdf_dtheta, dlogpdf_df_dtheta, d2logpdf_df2_dtheta