def testing(prefix, classifier):
feature, label, name = data.get_matrix('test', False)
predict = classifier.predict(feature)
res_file = open(prefix + '.txt', 'w')
res_file.write(str(classifier.score(feature, label) * 100.0) + '\n')
for i in xrange(4):
res_file.write(' '.join(map(str, (np.logical_and(label == i, predict == j).sum() for j in xrange(4)))) + '\n')
res_file.write('predict label name\n')
for x in zip(predict, label, name):
res_file.write(' '.join(map(str, x)) + '\n')
res_file.close()
def training(model_name):
feature, label, _ = data.get_matrix('train', True)
classifier = model.get_model(model_name)
classifier.fit(feature, label)
return classifier
def run():
group, name, index = parse_argv()
np_matrix = data.get_matrix(group, name, index)
matrix.run_operations_on_matrix(np_matrix)
from sklearn import manifold
import torch.nn.functional as F
import matplotlib.lines as mlines
import pandas as pd
import umap
import sys
#sys.path.append('../dataset')
#import graph_similarity_matrix as gsm
import data as mdata
#dpath1='/Users/iqbal/multiview3d/dataset_3D/clusters_dataset/dist_2.csv'
dpath1 = '/Users/iqbal/multiview3d/dataset_3D/123_dataset_new/250/data_mat_1_250.csv'
#dpath1='/Users/iqbal/multiview3d/dataset_3D/sq_cir_tr_dataset/350/data_mat_sq_350.csv'
D1 = mdata.get_matrix(dpath1)
#for i in range(1, 30):
# for j in range(1 )
pr = 20
ex = 12
lr = 1
filename = "input_squire_perplexity_" + str(pr)
#tsne = manifold.TSNE(n_components=3,perplexity=10.0, early_exaggeration=12.0, learning_rate=200.0, n_iter=1000, n_iter_without_progress=300, min_grad_norm=1e-07, metric='euclidean', init='random', verbose=1, random_state=None, method='barnes_hut', angle=0.5)
tsne = manifold.TSNE(n_components=2,
perplexity=pr,
early_exaggeration=ex,
learning_rate=lr,
n_iter=1000,
plt.savefig('precision_ml100k.png')
plt.show()
if __name__ == '__main__':
# # ml100k的数据
# nb_user=943
# nb_item=1682
# top_k = 5
# train_set_dict, test_set_dict = read_ml100k('dataset/ml-100k/u1.base', 'dataset/ml-100k/u1.test', sep='\t', header=None)
# train_set, test_set = get_matrix(train_set_dict, test_set_dict, nb_user=nb_user, nb_item=nb_item)
# train_CFGAN(train_set, nb_item, epoches=300, batch_size=32, nb_zr=128, nb_pm=128, alpha=0.1, test_set_dict=test_set_dict, top_k=top_k)
# ml1m的数据,超参数与ml100k不一样
nb_user = 6040
nb_item = 3952
top_k = 5
train_set_dict, test_set_dict = read_ml1m('dataset/ml-1m/ratings.dat')
train_set, test_set = get_matrix(train_set_dict,
test_set_dict,
nb_user=nb_user,
nb_item=nb_item)
train_CFGAN(train_set,
nb_item,
epoches=2000,
batch_size=128,
nb_zr=512,
nb_pm=512,
alpha=1,
test_set_dict=test_set_dict,
top_k=top_k)
diff = np.sum(np.square(vi - vj))
d = 0 if diff < eps else np.sqrt(diff)
cost = cost + np.square(d -
W[i][j]) if abs(d -
W[i][j]) > eps else cost
return cost
#dotpath='../dataset/game_of_thrones_consistent.dot'
#M, G, nodes_index=gsm.get_similarity_matrix(dotpath)
#X=np.zeros((6,2))
#X=np.random.rand(6,2)
#D1, D2, D3=data()
D1 = mdata.get_matrix('../dataset/dist_1.csv')
D2 = mdata.get_matrix('../dataset/dist_2.csv')
#D3=mdata.get_matrix('../dataset/dist_2.csv')
D3 = np.zeros((len(D1), len(D1)))
A = np.random.rand(len(D1) * dim, 1)
#B=A.copy()
#print(A)
pos1 = multiview_autograd(alpha, A, steps, dim, stopping_eps)
#pos2=mds_sklearn(alpha,A,steps,dim)
pos1 = pos1.reshape(int(len(pos1) / dim), dim)
#pos2=pos2.reshape(int(len(ZZ)/dim),dim)
fig = plt.figure()
ax = plt.axes(projection='3d')
# Updated for python3
# CS 251 Project 6
#
# PCA test function
#
import numpy as np
import data
import analysis
import sys
if __name__ == "__main__":
if len(sys.argv) < 2:
print('Usage: python %s <data file>' % (sys.argv[0]))
exit()
data = data.Data(sys.argv[1])
pcadata = analysis.pca(data, data.get_headers(), False)
print("\nOriginal Data Headers")
print(pcadata.get_original_headers())
print("\nOriginal Data")
print(data.get_matrix(data.get_headers()))
print("\nOriginal Data Means")
print(pcadata.get_original_means())
print("\nEigenvalues")
print(pcadata.get_eigenvalues())
print("\nEigenvectors")
print(pcadata.get_eigenvectors())
print("\nProjected Data")
print(pcadata.get_matrix(pcadata.get_headers()))
else:
d = np.sqrt(diff)
#print("diss",d)
if abs(d - M[i][j]) > eps:
cost = cost + np.square(d - M[i][j])
#dis[j][i]=dis[i][j]
#dis = euclidean_distances(X)
#cost1=np.square(dis-M)
#cost=np.sum(cost1)
return cost
#dotpath='../dataset/game_of_thrones_consistent.dot'
#M, G, nodes_index=gsm.get_similarity_matrix(dotpath)
M = mdata.get_matrix('../dataset/dist_2.csv')
M = M[0:29, 0:29]
print(M.shape)
#X=np.zeros((6,2))
#X=np.random.rand(6,2)
#M=data()
A = np.random.rand(len(M) * dim, 1)
#B=A.copy()
#print(A)
ZZ = mds_autograd(alpha, A, M, steps, dim)
Z = mds_sklearn(alpha, A, M, steps, dim)
fig = plt.figure()
if len(sys.argv) != 8:
print(
"uses: multiview_distance_matrix.py nameofdataset datapath1 datapath2 datapath3 learning_rate maxsteps"
)
sys.exit()
name_data_set = sys.argv[1]
dpath1 = sys.argv[2]
dpath2 = sys.argv[3]
dpath3 = sys.argv[4]
alpha = float(sys.argv[5])
steps = int(sys.argv[6])
outputpath = sys.argv[7]
D1 = mdata.get_matrix(dpath1)
D2 = mdata.get_matrix(dpath2)
D3 = mdata.get_matrix(dpath3)
#D1=D1[0:10,0:10]
#D2=D2[0:10,0:10]
#D3=D3[0:10,0:10]
P1 = np.random.rand(4, 1)
P2 = np.random.rand(4, 1)
P3 = np.random.rand(4, 1)
print("number of data points", len(D1))
A = np.random.rand(len(D1) * dim, 1)
mview = multiview(D1, D2, D3, dim, eps)
pos1, costs, P1, P2, P3 = mview.multiview_mds_projection(