def iterate(n_iters):
for i in tqdm(xrange(n_iters)):
sampler.sample()
likelihoods.append(sampler.tree.marg_log_likelihood())
plt.figure()
plt.xlabel("Iterations", fontsize=fontsize)
plt.ylabel("Data Log Likelihood", fontsize=fontsize)
plt.plot(likelihoods)
plt.legend(loc='best', fontsize=12)
plt.savefig('unconstrained-likelihoods.png', bbox_inches='tight')
final_tree = sampler.tree.copy()
plt.figure()
plot_tree_2d(final_tree, X, pca)
for node in final_tree.dfs():
if node.is_leaf():
node.point = y[node.point]
plt.figure()
newick = final_tree.to_newick()
tree = Phylo.read(StringIO(newick), 'newick')
Phylo.draw_graphviz(tree, prog='neato')
plt.savefig('unconstrained-tree.png', bbox_inches='tight')
graph = Phylo.to_networkx(tree)
with open('unconstrained-tree.nwk', 'w') as fp:
print >>fp, newick,
nx.write_dot(graph, 'unconstrained-tree.dot')
plt.show()
def iterate(n_iters):
for i in tqdm(xrange(n_iters)):
sampler.sample()
likelihoods.append(sampler.tree.marg_log_likelihood())
plt.figure()
plt.xlabel("Iterations", fontsize=fontsize)
plt.ylabel("Data Log Likelihood", fontsize=fontsize)
plt.plot(likelihoods)
plt.legend(loc='best', fontsize=12)
plt.savefig('unconstrained-likelihoods.png', bbox_inches='tight')
final_tree = sampler.tree.copy()
plt.figure()
plot_tree_2d(final_tree, X, pca)
for node in final_tree.dfs():
if node.is_leaf():
node.point = y[node.point]
plt.figure()
newick = final_tree.to_newick()
tree = Phylo.read(StringIO(newick), 'newick')
Phylo.draw_graphviz(tree, prog='neato')
plt.savefig('unconstrained-tree.png', bbox_inches='tight')
graph = Phylo.to_networkx(tree)
with open('unconstrained-tree.nwk', 'w') as fp:
print >> fp, newick,
nx.write_dot(graph, 'unconstrained-tree.dot')
plt.show()
from trees.util import plot_tree, plot_tree_2d
from trees.ddt import DirichletDiffusionTree, Inverse, GaussianLikelihoodModel
from trees.mcmc import MetropolisHastingsSampler
from tqdm import tqdm
if __name__ == "__main__":
D = 2
N = 100
X = np.random.multivariate_normal(mean=np.zeros(D), cov=np.eye(D), size=N).astype(np.float32)
df = Inverse(c=1)
lm = GaussianLikelihoodModel(sigma=np.eye(D) / 4.0, mu0=np.zeros(D), sigma0=np.eye(D))
ddt = DirichletDiffusionTree(df=df,
likelihood_model=lm)
mh = MetropolisHastingsSampler(ddt, X)
mh.initialize_assignments()
for _ in tqdm(xrange(1000)):
mh.sample()
plt.figure()
plt.plot(mh.likelihoods)
plt.figure()
plot_tree(mh.tree)
plt.figure()
plot_tree_2d(mh.tree, X)
plt.show()
plt.legend(loc='best', fontsize=12)
plt.savefig('online-scores.png', bbox_inches='tight')
plt.figure()
plt.xlim([0, n_iters + constraint_add])
plt.xlabel("Iterations", fontsize=fontsize)
plt.ylabel("Data Log Likelihood", fontsize=fontsize)
plt.plot(likelihoods)
plt.legend(loc='best', fontsize=12)
plt.savefig('online-likelihoods.png', bbox_inches='tight')
final_tree = sampler.tree.copy()
plt.figure()
plot_tree_2d(final_tree, X, pca)
for node in final_tree.dfs():
if node.is_leaf():
node.point = y[node.point]
newick = final_tree.to_newick()
tree = Phylo.read(StringIO(newick), 'newick')
plt.figure()
Phylo.draw_graphviz(tree, prog='neato')
plt.savefig('tree.png', bbox_inches='tight')
graph = Phylo.to_networkx(tree)
with open('tree.nwk', 'w') as fp:
print >> fp, newick,
nx.write_dot(graph, 'tree.dot')
plt.legend(loc='best', fontsize=12)
plt.savefig('offline-scores.png', bbox_inches='tight')
plt.figure()
plt.xlim([0, n_iters])
# plt.ylim(ymin=-400)
plt.xlabel("Iterations", fontsize=fontsize)
plt.ylabel("Data Log Likelihood", fontsize=fontsize)
for name, likelihood in likelihoods.items():
plt.plot(likelihood, label=name)
plt.legend(loc='best', fontsize=12)
plt.savefig('offline-likelihoods.png', bbox_inches='tight')
for type, model in models.items():
final_tree = model.copy()
plt.figure()
plot_tree_2d(final_tree, X, pca)
for node in final_tree.dfs():
if node.is_leaf():
node.point = y[node.point]
newick = final_tree.to_newick()
tree = Phylo.read(StringIO(newick), 'newick')
Phylo.draw_graphviz(tree, prog='neato')
plt.savefig('tree-%s.png' % type, bbox_inches='tight')
plt.show()
