dims_dg = [[256, 1024], [256, 1024]]
lr = [lr_pre, lr_ae, lr_dg, lr_h]
epochs = [epochs_pre, epochs_total, epochs_h]
for j in range(num):
data = Dataset('coil_2views')
x1, x2, gt = data.load_data()
x1 = data.normalize(x1, 0)
x2 = data.normalize(x2, 0)
X[str(0)], X[str(1)] = x1, x2
n_clusters = len(set(gt))
H, gt = model_multi_view(X, gt, para_lambda, dims_ae, dims_dg, act, lr,
epochs, batch_size)
acc_H_all[j], acc_H_std, nmi_H_all[j], nmi_H_std, RI_H_all[j], RI_std, f1_H_all[j], f1_std = \
cluster(n_clusters, H, gt, count=1)
print('clustering h : acc = {:.4f}, nmi = {:.4f}'.format(
acc_H_all[j], nmi_H_all[j]))
acc_mean = np.mean(acc_H_all)
nmi_mean = np.mean(nmi_H_all)
acc_std = np.std(acc_H_all)
nmi_std = np.std(nmi_H_all)
RI_mean = np.mean(RI_H_all)
RI_std = np.std(RI_H_all)
fs_mean = np.mean(f1_H_all)
fs_std = np.std(f1_H_all)
arg = [
def print_result(n_clusters, H, gt, count=10):
acc_avg,nmi_avg,ari_avg,f1_avg = cluster(n_clusters, H, gt, count=count)
print('clustering h : acc = {:.4f}, nmi = {:.4f},ari = {:.4f}, f1 = {:.4f}'.format(acc_avg, nmi_avg,ari_avg, f1_avg))
act = [act_ae1, act_ae2, act_dg1, act_dg2]
dims = [dims_ae1, dims_ae2, dims_dg1, dims_dg2]
lr = [lr_pre, lr_ae, lr_dg, lr_h]
epochs = [epochs_pre, epochs_total, epochs_h]
for j in range(num):
data = Dataset('ORL_2views')
x1, x2, gt = data.load_data()
x1 = data.normalize(x1, 0)
x2 = data.normalize(x2, 0)
n_clusters = len(set(gt))
H, gt = model(x1, x2, gt, para_lambda, dims, act, lr, epochs, batch_size)
acc_H_all[j], acc_H_std, nmi_H_all[j], nmi_H_std, RI_H_all[j], RI_std, f1_H_all[j], f1_std = cluster(n_clusters,
H, gt,
count=1)
print('clustering h : acc = {:.4f}, nmi = {:.4f}'.format(acc_H_all[j], nmi_H_all[j]))
acc_mean = np.mean(acc_H_all)
nmi_mean = np.mean(nmi_H_all)
acc_std = np.std(acc_H_all)
nmi_std = np.std(nmi_H_all)
RI_mean = np.mean(RI_H_all)
RI_std = np.std(RI_H_all)
fs_mean = np.mean(f1_H_all)
fs_std = np.std(f1_H_all)
arg = ['lambda', para_lambda, 'batch_size', batch_size, 'lr_pre', lr_pre, 'lr_ae', lr_ae,
dims = [dims_ae1, dims_ae2, dims_dg1, dims_dg2]
lr = [lr_pre, lr_ae, lr_dg, lr_h]
epochs = [epochs_pre, epochs_total, epochs_h]
for j in range(num):
data = Dataset('handwritten_2views')
x1, x2, gt = data.load_data()
x1 = data.normalize(x1, 0)
x2 = data.normalize(x2, 0)
n_clusters = len(set(gt))
H, gt = model(x1, x2, gt, para_lambda, dims, act, lr, epochs,
batch_size)
acc_H_all[j], acc_H_std, nmi_H_all[j], nmi_H_std, RI_H_all[
j], RI_std, f1_H_all[j], f1_std = cluster(n_clusters,
H,
gt,
count=1)
print('clustering h : acc = {:.4f}, nmi = {:.4f}'.format(
acc_H_all[j], nmi_H_all[j]))
acc_mean = np.mean(acc_H_all)
nmi_mean = np.mean(nmi_H_all)
acc_std = np.std(acc_H_all)
nmi_std = np.std(nmi_H_all)
RI_mean = np.mean(RI_H_all)
RI_std = np.std(RI_H_all)
fs_mean = np.mean(f1_H_all)
fs_std = np.std(f1_H_all)