def testGlass(loader, averaging=False):
target, data, nfeat, nsamples = loader.glass()
inputs = torch.from_numpy(data).unsqueeze(1).float()
ut = np.unique(target)
print(nsamples, 'samples,', nfeat, 'dimensions,targets:', Counter(target))
Nclusters = len(ut)
Nsparse = 40
batchsize = 40
Nfactor = 20
model = dmClustering(bioNet(dim=nfeat),
Nfactor=Nfactor,
Nclusters=Nclusters)
F, labels = model.unSupervisedLearner(inputs,
Nepochs=10,
Ninner=1,
sparsity=Nsparse,
bsize=batchsize,
lamda=1)
if not averaging:
print('clusters:', ''.join(str(l) for l in labels))
print(Counter(labels))
# print('targets:', ''.join(str(t) for t in target))
nmi = normalized_mutual_info_score(target, labels)
print('Normalized mutual information(NMI): {:.2f}%'.format(nmi * 100))
acc = group_label_acc(target, labels)
print('Group accuracy: {:.2f}%'.format(acc * 100))
return acc, nmi
else:
return group_label_acc(target, labels)
def testCol100(loader):
target, data, nfeat, nsamples = loader.col100()
inputs = torch.from_numpy(data).unsqueeze(1).float()
ut = np.unique(target)
print(nsamples, 'samples,', nfeat, 'dimensions,targets:', Counter(target))
Nclusters = len(ut)
Nsparse = 20
batchsize = 50
Nfactor = 10
model = dmClustering(objectNet1d(dim=nfeat),
Nfactor=Nfactor,
Nclusters=Nclusters)
F, labels = model.unSupervisedLearner(inputs,
Nepochs=10,
Ninner=1,
sparsity=Nsparse,
bsize=batchsize,
lamda=1)
ctr = Counter(labels)
print('clusters:', ''.join(str(l) for l in labels))
print(ctr)
nmi = normalized_mutual_info_score(target, labels)
print('Normalized mutual information(NMI): {:.2f}%'.format(nmi * 100))
acc = group_label_acc(target, labels)
print('Group accuracy: {:.2f}%'.format(acc * 100))
print('Number of clusters: {:d} '.format(len(ctr)))
return acc, nmi
def testSpiral(loader, averaging=False):
target, inputs, nfeat, nsamples = loader.spiral()
if type(inputs) is np.ndarray:
inputs = torch.from_numpy(inputs).unsqueeze(1).float()
ut = np.unique(target)
else:
ut = target.unique()
print(nsamples, ' samples', nfeat, ' dimensions ,targets:',
Counter(target))
Nclusters = len(ut)
Nsparse = 3
batchsize = 10
model = dmClustering(shapeNet(), Nfactor=0, Nclusters=Nclusters)
F, labels = model.unSupervisedLearner(inputs,
Nepochs=20,
sparsity=Nsparse,
bsize=batchsize,
lamda=1)
if not averaging:
#print(F)
print('clusters:', ''.join(str(l) for l in labels))
print(Counter(labels))
#print('targets:', ''.join(str(t) for t in target))
nmi = normalized_mutual_info_score(target, labels)
print('Normalized mutual information(NMI): {:.2f}%'.format(nmi * 100))
acc = group_label_acc(target, labels)
print('Group accuracy: {:.2f}%'.format(acc * 100))
return acc, nmi
else:
return group_label(target, labels)
def testUsps(loader, pretrained=True):
target, data, nfeat, nsamples = loader.usps()
inputs = torch.from_numpy(data).unsqueeze(1).float()
ut = np.unique(target)
print(nsamples, 'samples,', nfeat, 'dimensions,targets:', Counter(target))
Nclusters = len(ut)
modelFile = './USPS_metricNet.dat'
net = faceNet1d(dim=nfeat)
if pretrained:
net.load_state_dict(torch.load(modelFile))
model = dmClustering(net, Nclusters=Nclusters)
tau = 100000.0
dfx, _ = model.metricLearner(inputs,
torch.from_numpy(target),
Nepochs=1500,
bsize=1000,
tau=tau)
torch.save(net.state_dict(), modelFile)
labels = metricCluster(dfx, ratio=tau / 2)
ctr = Counter(labels)
print('clusters:', ''.join(str(l) for l in labels))
print(ctr)
nmi = normalized_mutual_info_score(target, labels)
print('Normalized mutual information(NMI): {:.2f}%'.format(nmi * 100))
acc = group_label_acc(target, labels)
print('Group accuracy: {:.2f}%'.format(acc * 100))
print('Number of clusters: {:d} '.format(len(ctr)))
return acc, nmi
def avgPred(ntimes, loader, testcase):
target, inputs, nfeat, nsamples = loader.spiral()
preds = torch.Tensor(ntimes, nsamples)
for i in range(ntimes):
print('predicting {:d}/{:d} round'.format(i + 1, ntimes))
if type(inputs) is np.ndarray:
preds[i, :] = torch.from_numpy(testcase(loader, True))
else:
preds[i, :] = testcase(loader, True)
print('max Voting')
pred, _ = preds.mode(0)
nmi = normalized_mutual_info_score(target, pred.numpy())
acc = group_label_acc(target, pred.numpy())
return acc, nmi
def testBlobs(tosave=False):
inputs, target = datasets.make_blobs(n_samples=600,
cluster_std=[2.0, 1.5, 1.0])
nsamples, nfeat = inputs.shape
inputsTensor = torch.from_numpy(inputs).unsqueeze(1).float()
ut = np.unique(target)
print(nsamples, ' samples', nfeat, ' dimensions ,targets:',
Counter(target))
Nclusters = len(ut)
Nsparse = 10
Nfactor = 0
batchsize = 25
model = dmClustering(shapeNet(dim=nfeat),
Nfactor=Nfactor,
Nclusters=Nclusters)
F, labels = model.unSupervisedLearner(inputsTensor,
Nepochs=6,
Ninner=1,
sparsity=Nsparse,
bsize=batchsize,
lamda=1)
print('clusters:', ''.join(str(l) for l in labels))
print(Counter(labels))
nmi = normalized_mutual_info_score(target, labels)
print('Normalized mutual information(NMI): {:.2f}%'.format(nmi * 100))
acc = group_label_acc(target, labels)
print('Group accuracy: {:.2f}%'.format(acc * 100))
model = SpectralClustering(n_clusters=2).fit(inputs)
labels_spcl = model.labels_
ss = torch.from_numpy(model.affinity_matrix_).float()
N = ss.size(0)
degs = ss.sum(1)
Ds = torch.eye(N)
Ds[Ds == 1] = degs
Ls = Ds - ss
_, v = Ls.symeig(eigenvectors=True)
f_spcl = v[:, :2].numpy()
fig, ax = plt.subplots(ncols=5, figsize=(20, 3))
c = ['b', 'r', 'm']
for i, l in enumerate(np.unique(target)):
id = (l == target)
ax[0].scatter(inputs[id, 0], inputs[id, 1], marker='.', color=c[i])
ax[0].set_title('Three Blobs')
for i, l in enumerate(np.unique(labels_spcl)):
id = (l == labels_spcl)
ax[1].scatter(inputs[id, 0], inputs[id, 1], marker='.', color=c[i])
ax[1].set_title('Spectral Clustering')
for i, l in enumerate(np.unique(labels)):
id = (l == labels)
ax[2].scatter(f_spcl[id, 0], f_spcl[id, 1], marker='.', color=c[i])
ax[2].set_title('Learnt Subspace')
for i, l in enumerate(np.unique(labels)):
id = (l == labels)
ax[3].scatter(inputs[id, 0], inputs[id, 1], marker='.', color=c[i])
ax[3].set_title('Proposed Clustering')
for i, l in enumerate(np.unique(labels)):
id = (l == labels)
ax[4].scatter(F.detach().numpy()[id, 0],
F.detach().numpy()[id, 1],
marker='.',
color=c[i])
ax[4].set_title('Learnt Subspace')
plt.tight_layout()
if tosave:
fig.savefig('ClusteringCompare-ThreeBlobs.png')
return 1
else:
plt.show(block=False)
time.sleep(5)
return 1
