def test___init__(self):
import numpy as np
f = ps.io.open(ps.examples.get_path('usjoin.csv'))
pci = np.array([f.by_col[str(y)] for y in range(1929, 2010)])
q5 = np.array([mc.Quantiles(y).yb for y in pci]).transpose()
m = Markov(q5)
np.testing.assert_array_almost_equal(markov_mobility(m.p, measure="P"),
0.19758992000997844)
def test___init__(self):
f = ps.io.open(ps.examples.get_path('usjoin.csv'))
pci = np.array([f.by_col[str(y)] for y in range(1929, 2010)])
q5 = np.array([mc.Quantiles(y).yb for y in pci]).transpose()
m = Markov(q5)
res = np.array(
[0.08988764, 0.21468144, 0.21125, 0.20194986, 0.07259074])
np.testing.assert_array_almost_equal(prais(m.p), res)
def G_display(G):
# make new graph
H = nx.Graph()
for v in G:
# print(v)
H.add_node(v)
weightValue = list(nx.get_edge_attributes(G, 'weight').values()) #提取权重
# weightsForWidth=[G[u][v]['weight'] for u,v in G.edges()] #another way
# print(weightValue)
import pysal.viz.mapclassify as mc
q = mc.Quantiles(weightValue, k=30).bins #计算分位数，用于显示值的提取
# print(q)
for (u, v, d) in tqdm(G.edges(data=True)):
# print(u,v,d)
# print()
# print(d['weight'])
if d['weight'] > q[28]:
H.add_edge(u, v)
print("H_digraph has %d nodes with %d edges" %
(nx.number_of_nodes(H), nx.number_of_edges(H)))
# draw with matplotlib/pylab
plt.figure(figsize=(18, 18))
# m=2
# fig = figure(figsize=(9*m,9*m)
# with nodes colored by degree sized by value elected
node_color = [float(H.degree(v)) for v in H]
# print(node_color)
# nx.draw(H, G.position,node_size=[G.perimeter[v] for v in H],node_color=node_color, with_labels=True)
weightsForWidthScale = np.interp(weightValue,
(min(weightValue), max(weightValue)),
(1, 3000)) #setting the edge width
scaleNode = 1
# sklearn.preprocessing.minmax_scale(X, feature_range=(0, 1), axis=0, copy=True)
nx.draw(H,
G.position,
node_size=minmax_scale([G.shape_area[v] * scaleNode for v in H],
feature_range=(10, 2200)),
node_color=node_color,
with_labels=True,
edge_cmap=plt.cm.Blues,
width=weightsForWidthScale) #edge_cmap=plt.cm.Blues
# scale the axes equally
# plt.xlim(-5000, 500)
# plt.ylim(-2000, 3500)
plt.show()
def test___init__(self):
pi = np.array([0.1, 0.2, 0.2, 0.4, 0.1])
f = ps.io.open(ps.examples.get_path('usjoin.csv'))
pci = np.array([f.by_col[str(y)] for y in range(1929, 2010)])
q5 = np.array([mc.Quantiles(y).yb for y in pci]).transpose()
m = Markov(q5)
np.testing.assert_array_almost_equal(markov_mobility(m.p, measure="P"),
0.19758992000997844)
np.testing.assert_array_almost_equal(markov_mobility(m.p, measure="D"),
0.60684854623695594)
np.testing.assert_array_almost_equal(
markov_mobility(m.p, measure="L2"), 0.039782002308159647)
np.testing.assert_array_almost_equal(
markov_mobility(m.p, measure="B1", ini=pi), 0.2277675878319787)
np.testing.assert_array_almost_equal(
markov_mobility(m.p, measure="B2", ini=pi), 0.046366601194789261)
def test___init__(self):
# markov = Markov(class_ids, classes)
f = ps.io.open(ps.examples.get_path('usjoin.csv'))
pci = np.array([f.by_col[str(y)] for y in range(1929, 2010)])
q5 = np.array([mc.Quantiles(y).yb for y in pci]).transpose()
m = Markov(q5)
expected = np.array([[729., 71., 1., 0., 0.], [72., 567., 80., 3., 0.],
[0., 81., 631., 86.,
2.], [0., 3., 86., 573., 56.],
[0., 0., 1., 57., 741.]])
np.testing.assert_array_equal(m.transitions, expected)
expected = np.array(
[[0.91011236, 0.0886392, 0.00124844, 0., 0.],
[0.09972299, 0.78531856, 0.11080332, 0.00415512, 0.],
[0., 0.10125, 0.78875, 0.1075, 0.0025],
[0., 0.00417827, 0.11977716, 0.79805014, 0.07799443],
[0., 0., 0.00125156, 0.07133917, 0.92740926]])
np.testing.assert_array_almost_equal(m.p, expected)
expected = np.array(
[0.20774716, 0.18725774, 0.20740537, 0.18821787, 0.20937187])
np.testing.assert_array_almost_equal(m.steady_state, expected)
alpha=1,
legend=True,
edgecolor='w',
linewidth=0.1)
plt.axis('equal')
plt.show()
# ### Quantiles
# One solution to obtain a more balanced classification scheme is using quantiles. This, by definition, assigns the same amount of values to each bin: the entire series is laid out in order and break points are assigned in a way that leaves exactly the same amount of observations between each of them. This "observation-based" approach contrasts with the "value-based" method of equal intervals and, although it can obscure the magnitude of extreme values, it can be more informative in cases with skewed distributions.
#
# Calculating a quantiles classification with `PySAL` can be done with the following line of code:
# In[38]:
classi = mapclassify.Quantiles(imd['imd_rank'], k=7)
classi
# And, similarly, the bins can also be inspected:
# In[39]:
classi.bins
# The visualization of the distribution can be generated in a similar way as well:
# In[40]:
# Set up the figure
f, ax = plt.subplots(1)
# Plot the kernel density estimation (KDE)
def networkAnalysis_B(G):
#本次实验中未使用该部分，但保留在该代码文件中，已备查看
try:
import pygraphviz # conda install -c alubbock pygraphviz 用此方法安装pygraphviz库，能正常安装
from networkx.drawing.nx_agraph import graphviz_layout
layout = graphviz_layout
except ImportError:
try:
import pydot
from networkx.drawing.nx_pydot import graphviz_layout
layout = graphviz_layout
except ImportError:
print("PyGraphviz and pydot not found;\n"
"drawing with spring layout;\n"
"will be slow.")
layout = nx.spring_layout
# 建立新网络图，复制G，make new graph
H = nx.Graph()
for v in G:
# print(v)
H.add_node(v)
weightValue = list(nx.get_edge_attributes(G, 'weight').values()) #提取权重值
# weightsForWidth=[G[u][v]['weight'] for u,v in G.edges()] #another way
# print(weightValue)
import pysal.viz.mapclassify as mc
q = mc.Quantiles(weightValue, k=50).bins #计算分位数，用于显示值的提取
# print(q)
#复制G到H
for (u, v, d) in G.edges(data=True):
# print(u,v,d)
# print()
# print(d['weight'])
if d['weight'] > q[48]: #根据权重选择较为紧密的部分提取
H.add_edge(u, v)
print("H_digraph has %d nodes with %d edges" %
(nx.number_of_nodes(H), nx.number_of_edges(H)))
weightsForWidthScale = np.interp(
weightValue, (min(weightValue), max(weightValue)),
(1, 3000)) #缩放值到新区间，用于强化显示数据，setting the edge width
# print(weightsForWidthScale)
node_color = [float(H.degree(v)) for v in H]
scaleNode = 1
# the following range of p values should be close to the threshold
plt.figure(figsize=(100, 100))
plt.subplots_adjust(left=0,
right=1,
bottom=0,
top=0.95,
wspace=0.01,
hspace=0.01)
# nx.draw(H, G.position, with_labels=False, node_size=800)
#打印图_node部分
nx.draw(H,
G.position,
node_size=minmax_scale([G.shape_area[v] * scaleNode for v in H],
feature_range=(100, 8200)),
node_color=node_color,
with_labels=True,
font_size=60,
edge_cmap=plt.cm.Blues,
width=weightsForWidthScale) #edge_cmap=plt.cm.Blues
#根据edge长度排序node identify largest connected component
Gcc = sorted(nx.connected_components(H), key=len, reverse=True)
# print(Gcc)
G0 = H.subgraph(Gcc[0]) #为最长的边
nx.draw_networkx_edges(G0,
G.position,
with_labels=False,
edge_color='r',
width=9.0)
#设置条件，显示其它长度的边 show other connected components
for Gi in Gcc[1:]:
if len(Gi) > 10:
nx.draw_networkx_edges(H.subgraph(Gi),
G.position,
with_labels=False,
edge_color='r',
alpha=0.3,
width=7.0)
plt.show()
