To make a graph using Networkx after spectral clustering on moons dataset

Question

I have generated moons dataset with 20 points and done spectral clustering on it. I want to form a graph using nearest neighbours = 3 with the help of Networkx. Where data points are the nodes and the affinity matrix generated after clustering is the weight over the edges between different nodes. I also need help in changing the colour and shape of the nodes of the two clusters so as to differentiate nodes of one cluster from another. Code is given below. An output image is given below. I just want to make a graph between the nodes of my output image using nearest neighbour=3.

import numpy as np
import os
from sklearn import metrics
from sklearn.cluster import SpectralClustering
from sklearn.neighbors import DistanceMetric
from sklearn.cluster import KMeans
import pandas as pd
import pylab as pl
import sklearn.metrics as sm
from sklearn.metrics import confusion_matrix,classification_report
from sklearn.preprocessing import MinMaxScaler
from sklearn.datasets import make_moons
import matplotlib.pyplot as plt
import networkx as nx
X, y = make_moons(n_samples=20)
print(X)
print(y)
plt.scatter(X[:,0],X[:,1], marker='o', facecolors='none', edgecolor='r')
clustering=SpectralClustering(n_clusters=2,
       assign_labels='kmeans',affinity='rbf',gamma=50, degree=3,
         random_state=0)
y_predict=clustering.fit_predict(X)
y_predict
clustering.labels_
clustering.affinity_matrix_
for i in range(0, y_predict.shape[0]):
    if y[i]==0 and y_predict[i]==0 :
        c1 = pl.scatter(X[i,0],X[i,1],c='b',
    marker='+')
    elif y[i]==1 and y_predict[i]==0:
        c2 = pl.scatter(X[i,0],X[i,1], facecolors='none', edgecolor='b',
    marker='o')
    elif y[i]==0 and y_predict[i]==1:
        c3=pl.scatter(X[i,0],X[i,1],c='r',
    marker='+')
    elif y[i]==1 and y_predict[i]==1:
        c4=pl.scatter(X[i,0],X[i,1], facecolors='none', edgecolor='r',
    marker='o')
pl.show()

Image of the clustering of moons dataset is given below

Answer 1

Building off of the answer to your previous question, this is what you asked for I believe.

Since the values in the affinity matrix are all between 0 and 1 but of very different relative magnitudes, I used -10 / log(weight) as the edge width.

import numpy as np
import os
from sklearn import metrics
from sklearn.cluster import SpectralClustering
from sklearn.neighbors import DistanceMetric
from sklearn.cluster import KMeans
import pandas as pd
import pylab as pl
import sklearn.metrics as sm
from sklearn.metrics import confusion_matrix,classification_report
from sklearn.preprocessing import MinMaxScaler
from sklearn.datasets import make_moons
import matplotlib.pyplot as plt
import networkx as nx
import math
X, y = make_moons(n_samples=20)
print(X)
print(y)
#plt.scatter(X[:,0],X[:,1], marker='o', facecolors='none', edgecolor='r')
pl.figure(figsize=(15, 12))
clustering=SpectralClustering(n_clusters=2,
       assign_labels='kmeans',affinity='rbf',gamma=50, degree=3,
         random_state=0)
y_predict=clustering.fit_predict(X)
y_predict
clustering.labels_
clustering.affinity_matrix_
for i in range(0, y_predict.shape[0]):
    if y[i]==0 and y_predict[i]==0 :
        c1 = pl.scatter(X[i,0],X[i,1],c='b',
    marker='+')
    elif y[i]==1 and y_predict[i]==0:
        c2 = pl.scatter(X[i,0],X[i,1], facecolors='none', edgecolor='b',
    marker='o')
    elif y[i]==0 and y_predict[i]==1:
        c3=pl.scatter(X[i,0],X[i,1],c='r',
    marker='+')
    elif y[i]==1 and y_predict[i]==1:
        c4=pl.scatter(X[i,0],X[i,1], facecolors='none', edgecolor='r',
    marker='o')
        
for i in range(0, len(X)):
  affinity_list = clustering.affinity_matrix_[i]
  affinity_list[i] = 0 # in case we don't want to consider the node as it's own neighbour
  nearest_neighbors_indices = np.argpartition(clustering.affinity_matrix_[i], -k)[-k:]
  for j in nearest_neighbors_indices:
    G.add_edge(tuple(X[i]), tuple(X[j]), weight = clustering.affinity_matrix_[i][j])

weights = [-10/math.log(edge[-1]['weight']) for edge in G.edges.data()]
# Draw Graph
pos = {node_name: node_name for node_name in G.nodes}
nx.draw_networkx_edges(G, pos, width=weights)
pl.show()