def test_MDS():
sim = np.array([[0, 5, 3, 4],
[5, 0, 2, 2],
[3, 2, 0, 1],
[4, 2, 1, 0]])
mds_clf = mds.MDS(metric=False, n_jobs=3, dissimilarity="precomputed")
mds_clf.fit(sim)
def isomap_own(X, no_of_dim, neighbour):
n = len(X)
adj = 1e10 * np.ones([n, n])
for i in range(len(adj)):
adj[i][i] = 0
for i in range(len(X)):
ndrs = NearestNeighbors(n_neighbors=neighbour,
algorithm='ball_tree',
metric='euclidean').fit(X)
dis, ind = ndrs.kneighbors([[X[i][0], X[i][1], X[i][2]]])
print(dis, ind)
print('**', len(dis))
for j in range(len(ind[0])):
adj[i][ind[0][j]] = dis[0][j]
adj[ind[0][j]][i] = dis[0][j]
print('****')
print(len(adj[0]))
print(adj[0])
dis = copy.deepcopy(adj)
n = len(adj)
for k in range(n):
for i in range(n):
for j in range(n):
if dis[i][j] > dis[i][k] + dis[k][j]:
dis[i][j] = dis[i][k] + dis[k][j]
dr = mds.MDS(n_components=2, dissimilarity="precomputed")
print(dr.fit_transform(dis))
y = dr.fit_transform(dis)
#return y
return mds_own(dis, no_of_dim)
def sklearn_mds(self):
mds_clf = mds.MDS(metric=True, n_jobs=3, dissimilarity="precomputed")
self.distanceMatrix = np.array(self.distanceMatrix)
#print self.distanceMatrix
res = mds_clf.fit_transform(self.distanceMatrix)
#print mds_clf
#print res
self.assign(res)
def test_MDS():
sim = np.array([[0, 5, 3, 4], [5, 0, 2, 2], [3, 2, 0, 1], [4, 2, 1, 0]])
mds_clf = mds.MDS(metric=False, n_jobs=3)
mds_clf.fit(sim)
dis = np.zeros([n_points, n_points])
X, color = datasets.samples_generator.make_s_curve(n_points, random_state=0)
print(X)
fig = plt.figure(figsize=(15, 8))
ax = fig.add_subplot(111, projection='3d')
ax.scatter(X[:, 0], X[:, 1], X[:, 2], c=color, cmap=plt.cm.Spectral)
ax.view_init(12, -72)
plt.show()
for i in range(n_points):
for j in range(n_points):
dis[i][j] = euclidean(X[i], X[j])
nx = mds_own(dis)
dr = mds.MDS(n_components=2)
nx1 = dr.fit_transform(X)
#nx=dr.fit_transform(X)
#print(nx)'''
'''dis=np.eye(len(X))
for i in range(len(X)):
for j in range(len(X)):
dis[i][j]=euclidean(X[i],X[j])
nx=mds_own(dis,2)'''
fig = plt.figure(figsize=(15, 8))
ax = fig.add_subplot(111)
ax.scatter(nx[:, 0], nx[:, 1], c=color, cmap=plt.cm.Spectral)
plt.show()