apunkt/GPy
1
0
Fork 0
mirror of https://github.com/SheffieldML/GPy.git synced 2026-06-08 15:05:15 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

xt

Browse source
This commit is contained in:
mu 2014-01-10 18:19:05 +00:00
parent 7072cfe7a0
commit 1aa060439f
1 changed files with 19 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

19
GPy/models/state_space_xt_sep.py
View file

@ -66,12 +66,31 @@ class StateSpace_1(Model):
return self.kern._get_param_names_transformed() + ['noise_variance'] return self.kern._get_param_names_transformed() + ['noise_variance']
def log_likelihood(self): def log_likelihood(self):
X=self.X
# Sort the matrix (save the order)
_, return_index, return_inverse = np.unique(X,True,True)
X = X[return_index]
# Get the model matrices from the kernel # Get the model matrices from the kernel
(F,L,Qc,H,Pinf) = self.kern.sde() (F,L,Qc,H,Pinf) = self.kern.sde()
n=X.shape[0]
F1 = np.kron(np.eye(n),F)
L1 = np.kron(np.eye(n),L)
K1=self.spacekern.K(X)
Qc1 = K1*Qc #kron(K,Qc1);
H1 = np.kron(np.eye(n),H)
Pinf1 = np.kron(K1,Pinf)
# Use the Kalman filter to evaluate the likelihood # Use the Kalman filter to evaluate the likelihood
return self.kf_likelihood(F,L,Qc,H,self.sigma2,Pinf,self.X.T,self.Y.T) return self.kf_likelihood(F,L,Qc,H,self.sigma2,Pinf,self.X.T,self.Y.T)
#return self.kf_likelihood(F1,L1,Qc1,H1,self.sigma2,Pinf1,self.X.T,self.Y.T)
def _log_likelihood_gradients(self): def _log_likelihood_gradients(self):