def test_log_likelihood_cn_partial_phi(self):
cn, h, l, phi, r, x = self.generate_simple_data()
emission = likelihood.NegBinLikelihood()
emission.h = h
emission.phi = phi
emission.r = r
def evaluate_log_likelihood(phi, x, l, cn):
emission.phi = phi
return np.sum(emission.log_likelihood(x, l, cn))
def evaluate_log_likelihood_partial_phi(phi, x, l, cn):
emission.phi = phi
return emission._log_likelihood_partial_phi(x, l, cn)[:, 0]
remixt.tests.utils.assert_grad_correct(
evaluate_log_likelihood, evaluate_log_likelihood_partial_phi, phi,
x, l, cn)
def test_learn_phi(self):
experiment = load_test_experiment()
emission = likelihood.NegBinLikelihood()
emission.h = experiment.h
emission.phi = experiment.phi
emission.r = experiment.negbin_r
N = experiment.l.shape[0]
M = experiment.h.shape[0]
prior = cn_model.CopyNumberPrior(N, M, perfect_cn_prior(experiment.cn))
prior.set_lengths(experiment.l)
model = cn_model.HiddenMarkovModel(N, M, emission, prior)
phi_init = experiment.phi + experiment.phi * 0.02 * np.random.randn(*experiment.phi.shape)
estimator = em.ExpectationMaximizationEstimator(num_em_iter=1)
estimator.learn_param(model, 'phi', phi_init)
def evaluate_log_likelihood_partial_h(h, x, l, cn, phi):
emission = likelihood.NegBinLikelihood()
emission.h = h
emission.phi = phi
emission.r = r
return emission._log_likelihood_partial_h(x, l, cn)