def test_pairwise_distance_matrix(get_covmats):
n_trials, n_channels = 6, 5
covmats = get_covmats(n_trials, n_channels)
n_subset = 4
A, B = covmats[:n_subset], covmats[n_subset:]
pdist = pairwise_distance(A, B)
assert pdist.shape == (n_subset, 2)
def _get_affinity_matrix(self, X, eps):
# make matrix with pairwise distances between points
distmatrix = pairwise_distance(X, metric=self.metric)
# determine which scale for the gaussian kernel
if self.eps is None:
eps = np.median(distmatrix)**2 / 2
# make kernel matrix from the distance matrix
kernel = np.exp(-distmatrix**2 / (4 * eps))
# normalize the kernel matrix
q = np.dot(kernel, np.ones(len(kernel)))
kernel_n = np.divide(kernel, np.outer(q, q))
return kernel_n
def test_pairwise_distance_matrix():
"""Test pairwise distance"""
A = np.array([2 * np.eye(3), 3 * np.eye(3)])
B = np.array([2 * np.eye(3), 3 * np.eye(3)])
pairwise_distance(A, B)
n_subjects = len(np.unique(np.array(metadata['subject'])))
##############################################################
# Aggregate all signals for a class from subjects
# and compute covmats. Then, compute dist between cov and
# pairwise distance
simi_vect, simi_mat = {}, {}
events = np.unique(y) # ['13', '17', '21', 'rest']
for e in events:
ev = aggregate_cov(X,
n_subjects,
metadata,
filterbank=True,
s_class=e,
estimator='lwf')
pdist = pairwise_distance(ev, metric=metric)
norm_fact = 1. / np.triu(pdist).sum()
simi = 1. / norm_fact * np.triu(pdist)
simi_mat[e] = simi
simi_vect[e] = simi[simi != 0.]
df = pd.DataFrame(simi_vect)
##############################################################
# Joint plot
for f in events[events != 'rest']:
g = sns.jointplot(x="rest", y=f, data=df)
plt.savefig('ssvep-jointplot-{0}-rest-{1}.png'.format(f, metric))
##############################################################
# Laplacian Eigenmaps
fig = plt.figure(figsize=(12, 3))
def test_pairwise_distance_matrix():
"""Test pairwise distance"""
A = np.array([2*np.eye(3), 3*np.eye(3)])
B = np.array([2*np.eye(3), 3*np.eye(3)])
pairwise_distance(A, B)