def draw_cdf(self,
sequence,
fig_ax=None,
title=None,
legend_label=None,
x_label="Degree",
y_label="CDF",
style='b-',
marker='o',
is_log=True):
if not fig_ax:
fig_ax = plt.subplots(figsize=figsize)
powerlaw.plot_cdf(sequence,
linewidth=linewidth,
marker=marker,
color=style[0],
ax=fig_ax[1],
label=legend_label)
fig_ax[1].grid(True)
if title:
fig_ax[1].set_title(title, fontsize=title_fontsize)
fig_ax[1].set_xlabel(x_label, fontsize=label_fontsize)
fig_ax[1].set_ylabel(y_label, fontsize=label_fontsize)
#fig_ax[1].set_xlim(left=0, right=500)
plt.yscale('linear')
if legend_label:
fig_ax[1].legend(fontsize=legend_fontsize)
#fig_ax[1].legend(fontsize=legend_fontsize, loc='lower right')
fig_ax[1].tick_params(size=tick_fontsize)
fig_ax[0].tight_layout()
return fig_ax
def plot_entropy_ccdf():
entropy = read_pickle('output/normalized_entropy.obj')
fig = plt.figure()
ax = fig.add_subplot(111)
powerlaw.plot_ccdf(entropy, ax, label='normalized entropy')
# further plotting
ax.set_xlabel("Normalized entropy e")
ax.set_ylabel("Pr(X>=e)")
plt.legend(fancybox=True, loc='lower left', ncol=1,prop={'size':5})
plt.tight_layout()
plt.savefig('output/normalized_entropy_distribution_ccdf.pdf')
fig = plt.figure()
ax = fig.add_subplot(111)
powerlaw.plot_cdf(entropy, ax, label='normalized entropy',color='r')
# further plotting
ax.set_xlabel("Normalized entropy e")
ax.set_ylabel("Pr(X<=e)")
plt.legend(fancybox=True, loc='lower left', ncol=1,prop={'size':5})
plt.tight_layout()
plt.savefig('output/normalized_entropy_distribution_cdf.pdf')
def plot_entropy_ccdf():
entropy = read_pickle('output/normalized_entropy.obj')
fig = plt.figure()
ax = fig.add_subplot(111)
powerlaw.plot_ccdf(entropy, ax, label='normalized entropy')
# further plotting
ax.set_xlabel("Normalized entropy e")
ax.set_ylabel("Pr(X>=e)")
plt.legend(fancybox=True, loc='lower left', ncol=1, prop={'size': 5})
plt.tight_layout()
plt.savefig('output/normalized_entropy_distribution_ccdf.pdf')
fig = plt.figure()
ax = fig.add_subplot(111)
powerlaw.plot_cdf(entropy, ax, label='normalized entropy', color='r')
# further plotting
ax.set_xlabel("Normalized entropy e")
ax.set_ylabel("Pr(X<=e)")
plt.legend(fancybox=True, loc='lower left', ncol=1, prop={'size': 5})
plt.tight_layout()
plt.savefig('output/normalized_entropy_distribution_cdf.pdf')
import scipy.io
import networkx as nx
import matplotlib.pyplot as plt
import powerlaw
sparse_mat = scipy.io.mmread('as-22july06.mtx')
G = nx.from_scipy_sparse_matrix(sparse_mat)
degree_sequence = sorted([y for x, y in G.degree()], reverse=True)
powerlaw.plot_cdf(degree_sequence)
# plt.show()
plt.savefig("cumulative-degree-distribution.png")
if __name__ == '__main__':
estimate_dir = "/home/valentin/Desktop/Thesis II/Zipf Error/Estimates"
lang = "NO"
estimate_file = lang + "_ToktokTokenizer_ArticleSplitter"
reader = TableReader(estimate_dir + "/" + estimate_file, [str, int, int])
data = reader.read_data()
counts = data["count"]
pos_counts = [c for c in counts if c > 0]
print(counts[:10])
print(min(counts))
powerlaw.plot_cdf(counts)
# plt.show()
powerlaw.plot_pdf(pos_counts)
# plt.show()
fitted_dist = powerlaw.Fit(pos_counts, discrete=True)
for key, val in fitted_dist.__dict__.items():
print(
key, ":\t", val
if hasattr(val, "__len__") and len(val) < 100 else "val too long")
print()
print("\n\n", fitted_dist.find_xmin())
r = df_primary.groupby(['patent_id'])['subgroup_id'].nunique()
z = r.value_counts().sort_index()
z.cumsum() / z.sum()
# In[12]:
df.groupby(['patent_id'])['subgroup_aggregated_id'].nunique().mean()
# In[13]:
df_primary.groupby(['patent_id'])['subgroup_aggregated_id'].nunique().mean()
# In[14]:
import powerlaw
powerlaw.plot_cdf(df_primary['group_id'].value_counts().values)
yscale('linear')
xscale('log')
print(df_primary['group_id'].nunique())
# In[15]:
import powerlaw
z = df_primary['subgroup_aggregated_id'].value_counts()
powerlaw.plot_cdf(z.values)
yscale('linear')
xscale('log')
print(len(z))
mean(z > sqrt(len(z)))**2
# In[16]:
def plot_features():
print 'before load'
network = load_graph("output/wikipedianetwork.xml.gz")
print 'after load'
print 'before load'
network_transitions = load_graph("output/transitionsnetwork.xml.gz")
print 'after load'
colors= {'local_clust':'r','eigenvector_centr':'b', 'page_rank': 'g', 'kcore':'m', 'hub': 'c', 'authority':'k'}
labels = {'local_clust': 'clust.', 'eigenvector_centr':'eigen. centr.','page_rank': 'page rank', 'kcore': 'kcore', 'hub':'hub', 'authority':'authority'}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust','page_rank', 'hub', 'authority', 'kcore']:
feature = network_transitions.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(), ax, label=labels[f],color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X>=f)$')
ax.set_ylim([0, 1])
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size':4})
plt.tight_layout()
plt.savefig('output/wikipedia-transitions-features-cdf.pdf')
plt.clf()
colors= {'local_clust':'r','eigenvector_centr':'b', 'page_rank': 'g', 'kcore':'m', 'hub': 'c', 'authority':'k'}
labels = {'local_clust': 'clust.', 'eigenvector_centr':'eigen. centr.','page_rank': 'page rank', 'kcore': 'kcore', 'hub':'hub', 'authority':'authority'}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust','page_rank', 'hub', 'authority', 'kcore']:
feature = network_transitions.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(), ax, label=labels[f],color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X<=f)$')
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size':4})
plt.tight_layout()
plt.savefig('output/wikipedia-transitions-features-ccdf.pdf')
plt.clf()
colors= {'local_clust':'r','eigenvector_centr':'b', 'page_rank': 'g', 'kcore':'m', 'hub': 'c', 'authority':'k'}
labels = {'local_clust': 'clust.', 'eigenvector_centr':'eigen. centr.','page_rank': 'page rank', 'kcore': 'kcore', 'hub':'hub', 'authority':'authority'}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust','page_rank', 'hub', 'authority', 'kcore']:
feature = network.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(), ax, label=labels[f],color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X>=f)$')
ax.set_ylim([0, 1])
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size':4})
plt.tight_layout()
plt.savefig('output/wikipedia-features-cdf.pdf')
plt.clf()
colors= {'local_clust':'r','eigenvector_centr':'b', 'page_rank': 'g', 'kcore':'m', 'hub': 'c', 'authority':'k'}
labels = {'local_clust': 'clust.', 'eigenvector_centr':'eigen. centr.','page_rank': 'page rank', 'kcore': 'kcore', 'hub':'hub', 'authority':'authority'}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust','eigenvector_centr','page_rank', 'hub', 'authority', 'kcore']:
feature = network.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(), ax, label=labels[f],color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X<=f)$')
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size':4})
plt.tight_layout()
plt.savefig('output/wikipedia-features-ccdf.pdf')
def plot_stats():
# wikipedia graph structural statistics
print 'before load'
network = load_graph("output/wikipedianetwork.xml.gz")
print 'after load'
out_hist = vertex_hist(network, "out")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(out_hist[1][:-1], out_hist[0], marker='o')
plt.xlabel('Out-degree')
plt.ylabel('Frequency')
plt.gca().set_ylim([1, 10**6])
#plt.title('Out-degree Distribution')
plt.tight_layout()
plt.savefig('output/wikipedia-out-deg-dist.pdf')
plt.clf()
in_hist = vertex_hist(network, "in")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(in_hist[1][:-1], in_hist[0], marker='o')
plt.xlabel('In-degree')
plt.ylabel('Frequency')
plt.gca().set_ylim([1, 10**6])
#plt.title('In-degree Distribution')
plt.tight_layout()
plt.savefig('output/wikipedia-in-deg-dist.pdf')
plt.clf()
total_hist = vertex_hist(network, "total")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(total_hist[1][:-1], total_hist[0], marker='o')
plt.xlabel('Degree')
plt.ylabel('Frequency')
plt.gca().set_ylim([1, 10**6])
#plt.title('Degree Distribution')
plt.tight_layout()
plt.savefig('output/wikipedia-deg-dist.pdf')
plt.clf()
clust = network.vertex_properties["local_clust"]
#clust = local_clustering(network, undirected=False)
#hist, bin_edges = np.histogram(clust.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Local Clustering Coefficient C')
#plt.ylabel('P(x<=C)')
#plt.title('Clustering Coefficient Distribution')
#plt.savefig('output/wikipedia-clust-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(clust.get_array(), ax)
#ax.set_title('Clustering Coefficient Distribution')
ax.set_xlabel('Local Clustering Coefficient $C')
ax.set_ylabel('P(x<=C)')
ax.set_ylim([0, 1])
fig.tight_layout()
fig.savefig('output/wikipedia-clust-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(clust.get_array(), ax)
#ax.set_title('Clustering Coefficient Distribution')
ax.set_xlabel('Local Clustering Coefficient C')
ax.set_ylabel('P(x>=C)')
ax.set_ylim([10**-4, 10**-0.5])
fig.tight_layout()
fig.savefig('output/wikipedia-clust-ccdf.pdf')
plt.clf()
prank = network.vertex_properties["page_rank"]
#hist, bin_edges = np.histogram(prank.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Page rank Pr')
#plt.ylabel('P(x<=Pr)')
#plt.title('Page rank Distribution')
#plt.savefig('output/wikipedia-prank-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(prank.get_array(), ax)
#ax.set_title('Page Rank Distribution')
ax.set_xlabel('Page rank Pr')
ax.set_ylabel('P(x<=Pr)')
ax.set_ylim([0, 1])
fig.tight_layout()
fig.savefig('output/wikipedia-prank-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(prank.get_array(), ax)
#ax.set_title('Page Rank Distribution')
ax.set_xlabel('Page rank Pr')
ax.set_ylabel('P(x>=Pr)')
fig.tight_layout()
fig.savefig('output/wikipedia-prank-ccdf.pdf')
plt.clf()
kcore = network.vertex_properties["kcore"]
#hist, bin_edges = np.histogram(kcore.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Kcore kC')
#plt.ylabel('P(x<=kC)')
#plt.title('K-Core Distribution')
#plt.savefig('output/wikipedia-kcore-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(kcore.get_array(), ax)
#ax.set_title('K-Core Distribution')
ax.set_xlabel('k-Core kC')
ax.set_ylabel('P(x<=kC)')
ax.set_ylim([0, 1])
fig.tight_layout()
fig.savefig('output/wikipedia-kcore-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(kcore.get_array(), ax)
#ax.set_title('K-Core Distribution')
ax.set_xlabel('k-Core kC')
ax.set_ylabel('P(x>=kC)')
fig.tight_layout()
fig.savefig('output/wikipedia-kcore-ccdf.pdf')
plt.clf()
eigenvector_centr = network.vertex_properties["eigenvector_centr"]
#hist, bin_edges = np.histogram(eigenvector_centr.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Eigenvector Centrality E')
#plt.ylabel('P(x<=E)')
#plt.title('Eigenvector Centrality Distribution')
#plt.savefig('output/wikipedia-eigenvcentr-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(eigenvector_centr.get_array(), ax)
#ax.set_title('Eigenvector Centrality E')
ax.set_xlabel('Eigenvector Centrality E')
ax.set_ylabel('P(x<=E)')
ax.set_ylim([0, 1])
fig.tight_layout()
fig.savefig('output/wikipedia-eigenvcentr-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(eigenvector_centr.get_array(), ax)
#ax.set_title('Eigenvector Centrality E')
ax.set_xlabel('Eigenvector Centrality E')
ax.set_ylabel('P(x>=E)')
fig.tight_layout()
fig.savefig('output/wikipedia-eigenvcentr-ccdf.pdf')
plt.clf()
colors= {'local_clust':'r','eigenvector_centr':'b', 'page_rank': 'g', 'kcore':'m', 'hub': 'c', 'authority':'k'}
labels = {'local_clust': 'clust.', 'eigenvector_centr':'eigen. centr.','page_rank': 'page rank', 'kcore': 'kcore', 'hub':'hub', 'authority':'authority'}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust','page_rank', 'hub', 'authority', 'kcore']:
feature = network.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(), ax, label=labels[f],color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X>=f)$')
ax.set_ylim([0, 1])
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size':4})
plt.tight_layout()
plt.savefig('output/wikipedia-features-cdf.pdf')
plt.clf()
colors= {'local_clust':'r','eigenvector_centr':'b', 'page_rank': 'g', 'kcore':'m', 'hub': 'c', 'authority':'k'}
labels = {'local_clust': 'clust.', 'eigenvector_centr':'eigen. centr.','page_rank': 'page rank', 'kcore': 'kcore', 'hub':'hub', 'authority':'authority'}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust','eigenvector_centr','page_rank', 'hub', 'authority', 'kcore']:
feature = network.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(), ax, label=labels[f],color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X<=f)$')
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size':4})
plt.tight_layout()
plt.savefig('output/wikipedia-features-ccdf.pdf')
plt.clf()
# wikipedia transitions graph structural statistics
print 'before load'
network_transitions = load_graph("output/transitionsnetwork.xml.gz")
print 'after load'
out_hist = vertex_hist(network_transitions, "out")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(out_hist[1][:-1], out_hist[0], marker='o')
plt.xlabel('Out-degree')
plt.ylabel('Frequency')
plt.gca().set_ylim([1, 10**6])
#plt.title('Out-degree Distribution')
plt.savefig('output/wikipedia-transitions-out-deg-dist.pdf')
plt.clf()
in_hist = vertex_hist(network_transitions, "in")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(in_hist[1][:-1], in_hist[0], marker='o')
plt.xlabel('In-degree')
plt.ylabel('Frequency')
#plt.title('In-degree Distribution')
plt.gca().set_ylim([1, 10**6])
plt.savefig('output/wikipedia-transitions-in-deg-dist.pdf')
plt.clf()
total_hist = vertex_hist(network_transitions, "total")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(total_hist[1][:-1], total_hist[0], marker='o')
plt.xlabel('Degree')
plt.ylabel('Frequency')
#plt.title('Degree Distribution')
plt.gca().set_ylim([1, 10**6])
plt.savefig('output/wikipedia-transitions-deg-dist.pdf')
plt.clf()
#clust = local_clustering(network_transitions, undirected=False)
clust = network_transitions.vertex_properties["local_clust"]
#hist, bin_edges = np.histogram(clust.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Local Clustering Coefficient C')
#plt.ylabel('P(x<=C)')
#plt.title('Clustering Coefficient Distribution')
#plt.savefig('output/wikipedia-transitions-clust-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(clust.get_array(), ax)
#ax.set_title('Clustering Coefficient Distribution')
ax.set_xlabel('Local Clustering Coefficient C')
ax.set_ylabel('P(x<=C)')
fig.savefig('output/wikipedia-transitions-clust-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(clust.get_array(), ax)
ax.set_title('Clustering Coefficient Distribution')
ax.set_xlabel('Local Clustering Coefficient C')
ax.set_ylabel('P(x>=C)')
fig.savefig('output/wikipedia-transitions-clust-ccdf.pdf')
plt.clf()
prank = network_transitions.vertex_properties["page_rank"]
#hist, bin_edges = np.histogram(prank.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Page rank Pr')
#plt.ylabel('P(x<=Pr)')
#plt.title('Page rank Distribution')
#plt.savefig('output/wikipedia-transitions-prank-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(prank.get_array(), ax)
#ax.set_title('Page Rank Distribution')
ax.set_xlabel('Page rank Pr')
ax.set_ylabel('P(x<=Pr)')
fig.savefig('output/wikipedia-transitions-prank-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(prank.get_array(), ax)
#ax.set_title('Page Rank Distribution')
ax.set_xlabel('Page rank Pr')
ax.set_ylabel('P(x>=Pr)')
fig.savefig('output/wikipedia-transitions-prank-ccdf.pdf')
plt.clf()
kcore = network_transitions.vertex_properties["kcore"]
#hist, bin_edges = np.histogram(kcore.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Kcore kC')
#plt.ylabel('P(x<=kC)')
#plt.title('K-Core Distribution')
#plt.savefig('output/wikipedia-transitions-kcore-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(kcore.get_array(), ax)
#ax.set_title('K-Core Distribution')
ax.set_xlabel('k-Core kC')
ax.set_ylabel('P(x<=kC)')
fig.savefig('output/wikipedia-transitions-kcore-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(kcore.get_array(), ax)
#ax.set_title('K-Core Distribution')
ax.set_xlabel('k-Core kC')
ax.set_ylabel('P(x>=kC)')
fig.savefig('output/wikipedia-transitions-kcore-ccdf.pdf')
plt.clf()
eigenvector_centr = network_transitions.vertex_properties["eigenvector_centr"]
#hist, bin_edges = np.histogram(eigenvector_centr.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Eingenvector centrality E')
#plt.ylabel('P(x<=E)')
#plt.title('Eigenvector Centrality Distribution')
#plt.savefig('output/wikipedia-transitions-eigenvcentr-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(eigenvector_centr.get_array(), ax)
#ax.set_title('Eigenvector Centrality Distribution')
ax.set_xlabel('Eingenvector centrality E')
ax.set_ylabel('P(x<=E)')
fig.savefig('output/wikipedia-transitions-eigenvcentr-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(eigenvector_centr.get_array(), ax)
#ax.set_title('Eigenvector Centrality Distribution')
ax.set_xlabel('Eingenvector centrality E')
ax.set_ylabel('P(x>=E)')
fig.savefig('output/wikipedia-transitions-eigenvcentr-ccdf.pdf')
plt.clf()
print 'before hits'
#ee, authority, hub = hits(network_transitions)
#network_transitions.vertex_properties["authority"] = authority
#network_transitions.vertex_properties["hub"] = hub
#network_transitions.save("output/transitionsnetwork.xml.gz")
print 'after hits'
colors= {'local_clust':'r','eigenvector_centr':'b', 'page_rank': 'g', 'kcore':'m', 'hub': 'c', 'authority':'k'}
labels = {'local_clust': 'clust.', 'eigenvector_centr':'eigen. centr.','page_rank': 'page rank', 'kcore': 'kcore', 'hub':'hub', 'authority':'authority'}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust','page_rank', 'hub', 'authority', 'kcore']:
feature = network_transitions.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(), ax, label=labels[f],color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X>=f)$')
ax.set_ylim([0, 1])
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size':4})
plt.tight_layout()
plt.savefig('output/wikipedia-transitions-features-cdf.pdf')
plt.clf()
colors= {'local_clust':'r','eigenvector_centr':'b', 'page_rank': 'g', 'kcore':'m', 'hub': 'c', 'authority':'k'}
labels = {'local_clust': 'clust.', 'eigenvector_centr':'eigen. centr.','page_rank': 'page rank', 'kcore': 'kcore', 'hub':'hub', 'authority':'authority'}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust','page_rank', 'hub', 'authority', 'kcore']:
feature = network_transitions.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(), ax, label=labels[f],color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X<=f)$')
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size':4})
plt.tight_layout()
plt.savefig('output/wikipedia-transitions-features-ccdf.pdf')
plt.clf()
def plot_features():
print 'before load'
network = load_graph("output/wikipedianetwork.xml.gz")
print 'after load'
print 'before load'
network_transitions = load_graph("output/transitionsnetwork.xml.gz")
print 'after load'
colors = {
'local_clust': 'r',
'eigenvector_centr': 'b',
'page_rank': 'g',
'kcore': 'm',
'hub': 'c',
'authority': 'k'
}
labels = {
'local_clust': 'clust.',
'eigenvector_centr': 'eigen. centr.',
'page_rank': 'page rank',
'kcore': 'kcore',
'hub': 'hub',
'authority': 'authority'
}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust', 'page_rank', 'hub', 'authority', 'kcore']:
feature = network_transitions.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(),
ax,
label=labels[f],
color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X>=f)$')
ax.set_ylim([0, 1])
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size': 4})
plt.tight_layout()
plt.savefig('output/wikipedia-transitions-features-cdf.pdf')
plt.clf()
colors = {
'local_clust': 'r',
'eigenvector_centr': 'b',
'page_rank': 'g',
'kcore': 'm',
'hub': 'c',
'authority': 'k'
}
labels = {
'local_clust': 'clust.',
'eigenvector_centr': 'eigen. centr.',
'page_rank': 'page rank',
'kcore': 'kcore',
'hub': 'hub',
'authority': 'authority'
}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust', 'page_rank', 'hub', 'authority', 'kcore']:
feature = network_transitions.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(),
ax,
label=labels[f],
color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X<=f)$')
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size': 4})
plt.tight_layout()
plt.savefig('output/wikipedia-transitions-features-ccdf.pdf')
plt.clf()
colors = {
'local_clust': 'r',
'eigenvector_centr': 'b',
'page_rank': 'g',
'kcore': 'm',
'hub': 'c',
'authority': 'k'
}
labels = {
'local_clust': 'clust.',
'eigenvector_centr': 'eigen. centr.',
'page_rank': 'page rank',
'kcore': 'kcore',
'hub': 'hub',
'authority': 'authority'
}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust', 'page_rank', 'hub', 'authority', 'kcore']:
feature = network.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(),
ax,
label=labels[f],
color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X>=f)$')
ax.set_ylim([0, 1])
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size': 4})
plt.tight_layout()
plt.savefig('output/wikipedia-features-cdf.pdf')
plt.clf()
colors = {
'local_clust': 'r',
'eigenvector_centr': 'b',
'page_rank': 'g',
'kcore': 'm',
'hub': 'c',
'authority': 'k'
}
labels = {
'local_clust': 'clust.',
'eigenvector_centr': 'eigen. centr.',
'page_rank': 'page rank',
'kcore': 'kcore',
'hub': 'hub',
'authority': 'authority'
}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in [
'local_clust', 'eigenvector_centr', 'page_rank', 'hub',
'authority', 'kcore'
]:
feature = network.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(),
ax,
label=labels[f],
color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X<=f)$')
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size': 4})
plt.tight_layout()
plt.savefig('output/wikipedia-features-ccdf.pdf')
def plot_stats():
# wikipedia graph structural statistics
print 'before load'
network = load_graph("output/wikipedianetwork.xml.gz")
print 'after load'
out_hist = vertex_hist(network, "out")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(out_hist[1][:-1], out_hist[0], marker='o')
plt.xlabel('Out-degree')
plt.ylabel('Frequency')
plt.gca().set_ylim([1, 10**6])
#plt.title('Out-degree Distribution')
plt.tight_layout()
plt.savefig('output/wikipedia-out-deg-dist.pdf')
plt.clf()
in_hist = vertex_hist(network, "in")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(in_hist[1][:-1], in_hist[0], marker='o')
plt.xlabel('In-degree')
plt.ylabel('Frequency')
plt.gca().set_ylim([1, 10**6])
#plt.title('In-degree Distribution')
plt.tight_layout()
plt.savefig('output/wikipedia-in-deg-dist.pdf')
plt.clf()
total_hist = vertex_hist(network, "total")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(total_hist[1][:-1], total_hist[0], marker='o')
plt.xlabel('Degree')
plt.ylabel('Frequency')
plt.gca().set_ylim([1, 10**6])
#plt.title('Degree Distribution')
plt.tight_layout()
plt.savefig('output/wikipedia-deg-dist.pdf')
plt.clf()
clust = network.vertex_properties["local_clust"]
#clust = local_clustering(network, undirected=False)
#hist, bin_edges = np.histogram(clust.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Local Clustering Coefficient C')
#plt.ylabel('P(x<=C)')
#plt.title('Clustering Coefficient Distribution')
#plt.savefig('output/wikipedia-clust-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(clust.get_array(), ax)
#ax.set_title('Clustering Coefficient Distribution')
ax.set_xlabel('Local Clustering Coefficient $C')
ax.set_ylabel('P(x<=C)')
ax.set_ylim([0, 1])
fig.tight_layout()
fig.savefig('output/wikipedia-clust-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(clust.get_array(), ax)
#ax.set_title('Clustering Coefficient Distribution')
ax.set_xlabel('Local Clustering Coefficient C')
ax.set_ylabel('P(x>=C)')
ax.set_ylim([10**-4, 10**-0.5])
fig.tight_layout()
fig.savefig('output/wikipedia-clust-ccdf.pdf')
plt.clf()
prank = network.vertex_properties["page_rank"]
#hist, bin_edges = np.histogram(prank.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Page rank Pr')
#plt.ylabel('P(x<=Pr)')
#plt.title('Page rank Distribution')
#plt.savefig('output/wikipedia-prank-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(prank.get_array(), ax)
#ax.set_title('Page Rank Distribution')
ax.set_xlabel('Page rank Pr')
ax.set_ylabel('P(x<=Pr)')
ax.set_ylim([0, 1])
fig.tight_layout()
fig.savefig('output/wikipedia-prank-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(prank.get_array(), ax)
#ax.set_title('Page Rank Distribution')
ax.set_xlabel('Page rank Pr')
ax.set_ylabel('P(x>=Pr)')
fig.tight_layout()
fig.savefig('output/wikipedia-prank-ccdf.pdf')
plt.clf()
kcore = network.vertex_properties["kcore"]
#hist, bin_edges = np.histogram(kcore.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Kcore kC')
#plt.ylabel('P(x<=kC)')
#plt.title('K-Core Distribution')
#plt.savefig('output/wikipedia-kcore-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(kcore.get_array(), ax)
#ax.set_title('K-Core Distribution')
ax.set_xlabel('k-Core kC')
ax.set_ylabel('P(x<=kC)')
ax.set_ylim([0, 1])
fig.tight_layout()
fig.savefig('output/wikipedia-kcore-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(kcore.get_array(), ax)
#ax.set_title('K-Core Distribution')
ax.set_xlabel('k-Core kC')
ax.set_ylabel('P(x>=kC)')
fig.tight_layout()
fig.savefig('output/wikipedia-kcore-ccdf.pdf')
plt.clf()
eigenvector_centr = network.vertex_properties["eigenvector_centr"]
#hist, bin_edges = np.histogram(eigenvector_centr.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Eigenvector Centrality E')
#plt.ylabel('P(x<=E)')
#plt.title('Eigenvector Centrality Distribution')
#plt.savefig('output/wikipedia-eigenvcentr-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(eigenvector_centr.get_array(), ax)
#ax.set_title('Eigenvector Centrality E')
ax.set_xlabel('Eigenvector Centrality E')
ax.set_ylabel('P(x<=E)')
ax.set_ylim([0, 1])
fig.tight_layout()
fig.savefig('output/wikipedia-eigenvcentr-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(eigenvector_centr.get_array(), ax)
#ax.set_title('Eigenvector Centrality E')
ax.set_xlabel('Eigenvector Centrality E')
ax.set_ylabel('P(x>=E)')
fig.tight_layout()
fig.savefig('output/wikipedia-eigenvcentr-ccdf.pdf')
plt.clf()
colors = {
'local_clust': 'r',
'eigenvector_centr': 'b',
'page_rank': 'g',
'kcore': 'm',
'hub': 'c',
'authority': 'k'
}
labels = {
'local_clust': 'clust.',
'eigenvector_centr': 'eigen. centr.',
'page_rank': 'page rank',
'kcore': 'kcore',
'hub': 'hub',
'authority': 'authority'
}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust', 'page_rank', 'hub', 'authority', 'kcore']:
feature = network.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(),
ax,
label=labels[f],
color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X>=f)$')
ax.set_ylim([0, 1])
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size': 4})
plt.tight_layout()
plt.savefig('output/wikipedia-features-cdf.pdf')
plt.clf()
colors = {
'local_clust': 'r',
'eigenvector_centr': 'b',
'page_rank': 'g',
'kcore': 'm',
'hub': 'c',
'authority': 'k'
}
labels = {
'local_clust': 'clust.',
'eigenvector_centr': 'eigen. centr.',
'page_rank': 'page rank',
'kcore': 'kcore',
'hub': 'hub',
'authority': 'authority'
}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in [
'local_clust', 'eigenvector_centr', 'page_rank', 'hub',
'authority', 'kcore'
]:
feature = network.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(),
ax,
label=labels[f],
color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X<=f)$')
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size': 4})
plt.tight_layout()
plt.savefig('output/wikipedia-features-ccdf.pdf')
plt.clf()
# wikipedia transitions graph structural statistics
print 'before load'
network_transitions = load_graph("output/transitionsnetwork.xml.gz")
print 'after load'
out_hist = vertex_hist(network_transitions, "out")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(out_hist[1][:-1], out_hist[0], marker='o')
plt.xlabel('Out-degree')
plt.ylabel('Frequency')
plt.gca().set_ylim([1, 10**6])
#plt.title('Out-degree Distribution')
plt.savefig('output/wikipedia-transitions-out-deg-dist.pdf')
plt.clf()
in_hist = vertex_hist(network_transitions, "in")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(in_hist[1][:-1], in_hist[0], marker='o')
plt.xlabel('In-degree')
plt.ylabel('Frequency')
#plt.title('In-degree Distribution')
plt.gca().set_ylim([1, 10**6])
plt.savefig('output/wikipedia-transitions-in-deg-dist.pdf')
plt.clf()
total_hist = vertex_hist(network_transitions, "total")
plt.gca().set_yscale('log')
plt.gca().set_xscale('log')
plt.plot(total_hist[1][:-1], total_hist[0], marker='o')
plt.xlabel('Degree')
plt.ylabel('Frequency')
#plt.title('Degree Distribution')
plt.gca().set_ylim([1, 10**6])
plt.savefig('output/wikipedia-transitions-deg-dist.pdf')
plt.clf()
#clust = local_clustering(network_transitions, undirected=False)
clust = network_transitions.vertex_properties["local_clust"]
#hist, bin_edges = np.histogram(clust.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Local Clustering Coefficient C')
#plt.ylabel('P(x<=C)')
#plt.title('Clustering Coefficient Distribution')
#plt.savefig('output/wikipedia-transitions-clust-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(clust.get_array(), ax)
#ax.set_title('Clustering Coefficient Distribution')
ax.set_xlabel('Local Clustering Coefficient C')
ax.set_ylabel('P(x<=C)')
fig.savefig('output/wikipedia-transitions-clust-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(clust.get_array(), ax)
ax.set_title('Clustering Coefficient Distribution')
ax.set_xlabel('Local Clustering Coefficient C')
ax.set_ylabel('P(x>=C)')
fig.savefig('output/wikipedia-transitions-clust-ccdf.pdf')
plt.clf()
prank = network_transitions.vertex_properties["page_rank"]
#hist, bin_edges = np.histogram(prank.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Page rank Pr')
#plt.ylabel('P(x<=Pr)')
#plt.title('Page rank Distribution')
#plt.savefig('output/wikipedia-transitions-prank-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(prank.get_array(), ax)
#ax.set_title('Page Rank Distribution')
ax.set_xlabel('Page rank Pr')
ax.set_ylabel('P(x<=Pr)')
fig.savefig('output/wikipedia-transitions-prank-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(prank.get_array(), ax)
#ax.set_title('Page Rank Distribution')
ax.set_xlabel('Page rank Pr')
ax.set_ylabel('P(x>=Pr)')
fig.savefig('output/wikipedia-transitions-prank-ccdf.pdf')
plt.clf()
kcore = network_transitions.vertex_properties["kcore"]
#hist, bin_edges = np.histogram(kcore.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Kcore kC')
#plt.ylabel('P(x<=kC)')
#plt.title('K-Core Distribution')
#plt.savefig('output/wikipedia-transitions-kcore-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(kcore.get_array(), ax)
#ax.set_title('K-Core Distribution')
ax.set_xlabel('k-Core kC')
ax.set_ylabel('P(x<=kC)')
fig.savefig('output/wikipedia-transitions-kcore-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(kcore.get_array(), ax)
#ax.set_title('K-Core Distribution')
ax.set_xlabel('k-Core kC')
ax.set_ylabel('P(x>=kC)')
fig.savefig('output/wikipedia-transitions-kcore-ccdf.pdf')
plt.clf()
eigenvector_centr = network_transitions.vertex_properties[
"eigenvector_centr"]
#hist, bin_edges = np.histogram(eigenvector_centr.get_array(), 100, density=True)
#cdf = np.cumsum(hist)
#plt.plot(bin_edges[1:], cdf, marker='o')
#plt.xlabel('Eingenvector centrality E')
#plt.ylabel('P(x<=E)')
#plt.title('Eigenvector Centrality Distribution')
#plt.savefig('output/wikipedia-transitions-eigenvcentr-cdf.pdf')
fig, ax = plt.subplots()
powerlaw.plot_cdf(eigenvector_centr.get_array(), ax)
#ax.set_title('Eigenvector Centrality Distribution')
ax.set_xlabel('Eingenvector centrality E')
ax.set_ylabel('P(x<=E)')
fig.savefig('output/wikipedia-transitions-eigenvcentr-cdf.pdf')
plt.clf()
fig, ax = plt.subplots()
powerlaw.plot_ccdf(eigenvector_centr.get_array(), ax)
#ax.set_title('Eigenvector Centrality Distribution')
ax.set_xlabel('Eingenvector centrality E')
ax.set_ylabel('P(x>=E)')
fig.savefig('output/wikipedia-transitions-eigenvcentr-ccdf.pdf')
plt.clf()
print 'before hits'
#ee, authority, hub = hits(network_transitions)
#network_transitions.vertex_properties["authority"] = authority
#network_transitions.vertex_properties["hub"] = hub
#network_transitions.save("output/transitionsnetwork.xml.gz")
print 'after hits'
colors = {
'local_clust': 'r',
'eigenvector_centr': 'b',
'page_rank': 'g',
'kcore': 'm',
'hub': 'c',
'authority': 'k'
}
labels = {
'local_clust': 'clust.',
'eigenvector_centr': 'eigen. centr.',
'page_rank': 'page rank',
'kcore': 'kcore',
'hub': 'hub',
'authority': 'authority'
}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust', 'page_rank', 'hub', 'authority', 'kcore']:
feature = network_transitions.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(),
ax,
label=labels[f],
color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X>=f)$')
ax.set_ylim([0, 1])
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size': 4})
plt.tight_layout()
plt.savefig('output/wikipedia-transitions-features-cdf.pdf')
plt.clf()
colors = {
'local_clust': 'r',
'eigenvector_centr': 'b',
'page_rank': 'g',
'kcore': 'm',
'hub': 'c',
'authority': 'k'
}
labels = {
'local_clust': 'clust.',
'eigenvector_centr': 'eigen. centr.',
'page_rank': 'page rank',
'kcore': 'kcore',
'hub': 'hub',
'authority': 'authority'
}
fig = plt.figure()
ax = fig.add_subplot(111)
for f in ['local_clust', 'page_rank', 'hub', 'authority', 'kcore']:
feature = network_transitions.vertex_properties[f]
powerlaw.plot_cdf(feature.get_array(),
ax,
label=labels[f],
color=colors[f])
ax.set_xlabel('Feature $f$')
ax.set_ylabel('$P(X<=f)$')
plt.legend(fancybox=True, loc=3, ncol=2, prop={'size': 4})
plt.tight_layout()
plt.savefig('output/wikipedia-transitions-features-ccdf.pdf')
plt.clf()
avalanchetoolbox.avalanches.signal_variability(d_all[33,:],)
pyplot.xlabel('Signal range (Standard deviation)')
pyplot.ylabel('Range probability (log(p))')
plots.savefig()
pyplot.close('all')
print('Analyzing all channels')
active_sensors = avalanchetoolbox.avalanches.signal_variability(d_all, (8,8))
plots.savefig()
pyplot.close('all')
print('Running avalanche analyses')
avs = array([])
for i in range(80):
d = mat[data_key][:64,:,i]
m = avalanchetoolbox.avalanches.run_analysis(d, time_scale='mean_iei', threshold_mode='Likelihood', threshold_level=10)
if 'size_events' in m.keys():
avs = concatenate((avs, m['size_events']), 1)
pyplot.figure()
powerlaw.plot_cdf(avs)
pyplot.xlim(1,100)
pyplot.plot((active_sensors, active_sensors), pyplot.ylim())
pyplot.title('Neuronal avalanche size distribution, survival function')
pyplot.xlabel('Avalanche Size (number of events)')
pyplot.ylabel('P(Size>x)')
plots.savefig()
pyplot.close('all')
plots.close()
#Used powerlaw package: https://github.com/jeffalstott/powerlaw
import networkx as nx
import matplotlib.pyplot as plt
import powerlaw as pl
#The graph is read as a weighted edgelist
G = nx.Graph()
G = nx.read_weighted_edgelist('as-22july06.mtx')
#Each node in the dataset has a corresponding degree - deg references to each
#node and iterates through the degrees of the graph and sorts it
sorted_degree = sorted([deg for node, deg in G.degree()])
#cdf - cumulative distribution function is a function under the powerlaw function
pl.cdf(data=sorted_degree, survival=False)
# plot_cdf function plots the cdf - also under the powerlaw package
pl.plot_cdf(sorted_degree)
plt.show()