def predict(self, Xstar, return_std=False, nsamples=10):
"""
Returns mean and covariances for each posterior sampled Student-t Process.
Parameters
----------
Xstar: np.ndarray, shape=((nsamples, nfeatures))
Testing instances to predict.
return_std: bool
Whether to return the standard deviation of the posterior process. Otherwise,
it returns the whole covariance matrix of the posterior process.
nsamples:
Number of posterior MCMC samples to consider.
Returns
-------
np.ndarray
Mean of the posterior process for each MCMC sample and Xstar.
np.ndarray
Covariance posterior process for each MCMC sample and Xstar.
"""
chunk = list(self.trace)
chunk = chunk[::-1][:nsamples]
post_mean = []
post_var = []
for posterior_sample in chunk:
params = self._extractParam(posterior_sample,
self.covfunc.parameters)
covfunc = self.covfunc.__class__(**params)
gp = tStudentProcess(covfunc, nu=self.nu + self.n)
gp.fit(self.X, self.y)
m, s = gp.predict(Xstar, return_std=return_std)
post_mean.append(m)
post_var.append(s)
return np.array(post_mean), np.array(post_var)
def test_tSP():
rng = np.random.RandomState(0)
X = rng.uniform(0, 5, 20)[:, np.newaxis]
y = 0.5 * np.sin(3 * X[:, 0]) + rng.normal(0, 0.5, X.shape[0])
sexp = squaredExponential()
tsp = tStudentProcess(sexp)
tsp.fit(X, y)
def test_tSP_opt_nograd():
rng = np.random.RandomState(0)
X = rng.uniform(0, 5, 20)[:, np.newaxis]
y = 0.5 * np.sin(3 * X[:, 0]) + rng.normal(0, 0.5, X.shape[0])
sexp = squaredExponential()
tsp = tStudentProcess(sexp, optimize=True)
tsp.fit(X, y)
params = tsp.getcovparams()
assert 0.3 < params['l'] < 0.5
assert 0.3 < params['sigmaf'] < 0.6
assert 0.2 < params['sigman'] < 0.4