def example23():
attributes = attributes2
families = families2
S = setup.connections(attributes_list=attributes,families_list=families)
D = distances.dmatrices(S,input_type='similarities',
connect_components=True,connect_factor=1.5)
K = len(S); N = len(S[0])
p = perspective.Persp()
p.fix_Q(number=K, special='standard')
a,b,c=compare.all(D,p,title='florence marriage + business',
names=attributes, edges=S, verbose=1)
# distance to family with maximum number of marriage & loan links
colors = [D[0,-2],D[1,1]]
compare.plot(a,b,c,title='florence marriage + loan / correct colors',
names=attributes, edges=S, colors=colors, verbose=1)
# distance to family with maximum number of marriage only
#colors = [D[0,-2],D[0,-2]]
#compare.plot(a,b,c,title='florence marriage + business',
# names=attributes, edges=S, colors=colors, verbose=1)
# distance to family with maximum number of marriage only
S0 = setup.connections(attributes_list=['business'],families_list=families)
D = distances.dmatrices(S0,input_type='similarities',
connect_components=True,connect_factor=1.5)
colors = [D[0,3],D[0,3]]
compare.plot(a,b,c,title='florence marriage + loan / business colors',
names=attributes, edges=S, colors=colors, verbose=1)
def example3a():
attributes = attributes2
families = families2
S = setup.connections(attributes_list=attributes,families_list=families)
D = []
edges = []
for s in S:
edges.append(multigraph.from_matrix(s,transformation='reciprocal')\
['edges'])
d = multigraph.sim2dist(s)
def f(x):
if x <3:
y = x
else:
y = None
return y
d = multigraph.from_matrix(d,transformation=f)
D.append(d)
K = len(S); N = len(S[0])
p = perspective.Persp()
p.fix_Q(number=K, special='standard')
mv = mpse.MPSE(D,persp=p,verbose=1)
mv.setup_visualization(visualization='mds')
mv.initialize_X()
mv.optimize_X(verbose=1)
mv.figureX()
mv.figureY(edges=edges)
mv.figureH()
plt.show()
def example1():
attribute = 'marriage'
families = setup.find_families([attribute])
S = setup.connections([attribute],families)[0]
D_all = multigraph.sim2dist(S)
colors = D_all[-2]
D0 = multigraph.from_matrix(S,transformation='reciprocal')
edges = D0['edges']
mv = mds2.MDS(D0,verbose=1)
mv.initialize()
mv.stochastic()
mv.agd(min_step=1e-6)
mv.figureX(edges=True,colors=colors)
mv.figureH()
def f(x):
if x <3:
y = x
else:
y = None
return y
D = multigraph.from_matrix(D_all,transformation=f)
mv = mds2.MDS(D,verbose=1)
mv.initialize()
mv.stochastic()
mv.agd(min_step=1e-6)
mv.figureX(edges=edges,colors=colors)
mv.figureH()
def f(x):
if x <5:
y = x
else:
y = None
return y
D = multigraph.from_matrix(D_all,transformation=f)
mv = mds2.MDS(D,verbose=1)
mv.initialize()
mv.stochastic()
mv.agd(min_step=1e-6)
mv.figureX(edges=edges,colors=colors)
mv.figureH()
plt.show()
def compute_mds(num=2):
if num==2:
attributes = attributes2
families = families2
elif num==3:
attributes = attributes3
families = families3
S = setup.connections(attributes_list=attributes,families_list=families)
D = distances.dmatrices(S,input_type='similarities',
connect_components=True,connect_factor=1.5)
K = len(S); N = len(S[0])
fig, axs = plt.subplots(1,K,sharey=True,sharex=True)
for i in range(K):
vis = mds.MDS(D[i],labels=families)
vis.initialize()
vis.optimize(algorithm='agd')
vis.graph(title=attributes[i])
axs[i].title.set_text(attributes[i])
if i ==0 :
for n in range(N):
axs[i].scatter(vis.X[n,0],vis.X[n,1],label=families[n])
else:
for n in range(N):
axs[i].scatter(vis.X[n,0],vis.X[n,1])
fig.legend(loc=7)
fig.tight_layout()
fig.subplots_adjust(right=0.85)
plt.show(block=False)
fig = vis.figure(); plt.show(block=False)
#proj = perspective.Persp()
#proj.fix_Q(number=K, special='standard')
#mv = multiview.Multiview(D,persp=proj,labels=families)
#mv.setup_visualization()
#mv.initialize_X()
#mv.optimize_X(rate=0.002,max_iters=200)
#mv.figureX(); mv.figureY();
#mv.figure();
#mv.graphY(k=0); mv.graphY(k=1)
plt.show()
def example2():
attributes = attributes2
families = families2
S = setup.connections(attributes_list=attributes,families_list=families)
D1 = dissimilarities.from_matrix(S[0],transformation='reciprocal')
D2 = dissimilarities.from_matrix(S[1],transformation='reciprocal')
D = [D1,D2]
#D = distances.dmatrices(S,input_type='similarities',
# connect_components=True,connect_factor=1.5)
K = len(S); N = len(S[0])
p = perspective.Persp()
p.fix_Q(number=K, special='standard')
a,b,c=compare.all(D,p,title='florence marriage + business',
names=attributes, edges=S, verbose=1)
# distance to family with maximum number of marriage & loan links
colors = [D[0,-2],D[1,1]]
compare.plot(a,b,c,title='florence marriage + loan / correct colors',
names=attributes, edges=S, colors=colors, verbose=1)