def get_train_data():
train_x_y = []
global flag0_count
global flag1_count
flag0_count = 0
flag1_count = 0
sess = tf.InteractiveSession()
for i in range(1, 10):
file_path_community = './0814data/%d/community-standard.txt' % i
file_path_network = './0814data/%d/network.txt' % i
# print(file_path_network)
social_list = cnn_socialNet_read_data.get_standard_network(
file_path_community)
my_graph = cnn_socialNet_read_data.get_graph(file_path_network)
cnn_socialNet_read_data.add_flag_graph(my_graph, social_list)
edges = []
for (u, v, flag) in my_graph.edges.data('flag'):
# print(u, v, flag)
if int(flag) == 0:
flag0_count = flag0_count + 1
else:
flag1_count = flag1_count + 1
edges.append(((u, v), flag))
for j in range(len(edges)):
matrix1, row1, clown1 = cnn_socialNet_deal_data.get_jump1_3dimension_different_size_matrix(
my_graph, edges[j][0])
image1 = tf.convert_to_tensor(matrix1)
image1 = tf.image.convert_image_dtype(image1, tf.float32)
resize_image1 = tf.image.resize_images(image1, [128, 128],
method=3)
img_numpy1 = resize_image1.eval(session=sess)
matrix2, row2, clown2 = cnn_socialNet_deal_data.get_jump2_3dimension_different_size_matrix(
my_graph, edges[j][0])
image2 = tf.convert_to_tensor(matrix2)
image2 = tf.image.convert_image_dtype(image2, tf.float32)
resize_image2 = tf.image.resize_images(image2, [128, 128],
method=3)
img_numpy2 = resize_image2.eval(session=sess)
# print('resize_iamge', img_numpy)
# matrix1 = tf.constant(resize_image).eval()
# print(edges[j][1])
if int(edges[j][1]) == 1:
label = [1, 0]
else:
label = [0, 1]
train_x_y.append((img_numpy1, img_numpy2, label))
sess.close()
return train_x_y, flag0_count, flag1_count
def get_train_data():
train_x_y = []
global flag0_count
global flag1_count
flag0_count = 0
flag1_count = 0
sess = tf.InteractiveSession()
file_path_community = './small_data/karate-standard.txt'
file_path_network = './small_data/karate-edges.txt'
# print(file_path_network)
social_list = cnn_socialNet_read_data.get_standard_network(
file_path_community)
print(social_list)
my_graph = cnn_socialNet_read_data.get_graph(file_path_network, split=',')
print(my_graph.edges)
cnn_socialNet_read_data.add_flag_graph(my_graph, social_list)
edges = []
for (u, v, flag) in my_graph.edges.data('flag'):
print(u, v, flag)
if int(flag) == 0:
flag0_count = flag0_count + 1
else:
flag1_count = flag1_count + 1
edges.append(((u, v), flag))
for j in range(len(edges)):
matrix, row, clown = cnn_socialNet_deal_data.get_jump1_3dimension_different_size_matrix(
my_graph, edges[j][0])
image = tf.convert_to_tensor(matrix)
image = tf.image.convert_image_dtype(image, tf.float32)
resize_image = tf.image.resize_images(image, [128, 128], method=3)
img_numpy = resize_image.eval(session=sess)
# print('resize_iamge', img_numpy)
# matrix1 = tf.constant(resize_image).eval()
# print(edges[j][1])
if int(edges[j][1]) == 1:
label = [1, 0]
else:
label = [0, 1]
train_x_y.append((img_numpy, label))
sess.close()
return train_x_y, flag0_count, flag1_count
import networkx as nx
import cnn_socialNet_read_data
file_path = './0814data/1/community-standard.txt'
test_file_path = './0814data/test_community.txt'
social_list = cnn_socialNet_read_data.get_standard_network(test_file_path)
print(social_list)
# 测试获取网络
test_network = './0814data/test_nodes.txt'
network_file_path = '0814data/1/network.txt'
test_G = cnn_socialNet_read_data.get_graph(test_network)
print(test_G.edges())
# 测试为网络中每条边添加flag属性
edges = []
cnn_socialNet_read_data.add_flag_graph(test_G, social_list)
for (u, v, flag) in test_G.edges.data('flag'):
print(u, v, flag)
edges.append(((u, v), int(flag)))
# 划分社区算法
visited = {}
def breadth_first_search(root=None):
queue = []
social = []
nodes = test_G.nodes
def bfs(first_node):
order = [first_node]
A = np.array([12,15,19,21,26,33,38,41,42,45,48,55,58,59,60,61,71,76,80,82,83,89,102,104,109,110,117])
B = np.array([1,193,121,48,58,41,104,82,55,26,109,117,60,102,42,83,110,71,89,33,38,80,21,76,59,61,15,45])
# A = np.array([3,5,6,7],[1,2])
# B = np.array([3,5,7],[1,2,6])
# print('NMI:', NMI(A, B))
print('P:', precision(A, B))
print('R:', recall(A, B))
print('F1:',f1(A, B))
community1 = [['1','2','3','4','5','6','7'], ['8','9','10','11','12','13'], ['14','15','16','17','18']]
community2 = [['1', '2', '3', '4', '5', '7'], ['8', '9', '10', '11', '12', '13'],
['14', '15', '16', '17', '18','6']]
community3 = [['1', '2', '3', '4', '5', '7'], ['8', '9', '10', '11', '13'],
['14', '15', '16', '17', '18', '6', '12']]
test_network = './0814data/test_nodes.txt'
# network_file_path = '0814data/1/network.txt'
test_G = cnn_socialNet_read_data.get_graph(test_network)
print('Q:',Q1(community1, test_G))
print('Q:', Q1(community2, test_G))
print('Q:', Q1(community3, test_G))
karate_comm = [['1', '2', '3', '4', '5', '6', '7', '8', '11', '12', '13', '14', '17', '18', '20', '22'],
['9', '10', '15', '16', '19', '21', '23', '24', '25', '26', '27', '28', '29', '30', '31', '32', '33', '34']]
karate_network = './small_data/karate-edges.txt'
karate_G = cnn_socialNet_read_data.get_graph(karate_network,split=',')
print('karate_Q:',Q1(karate_comm,karate_G))
comm1 = ['1', '2', '3', '4', '5', '6', '7'] # 13/(13+2) = 0.86
comm2 = ['8', '9', '10', '11', '12', '13'] # 9/(9+2) = 0.81
comm3 = ['14', '15', '16', '17', '18'] # 8/(8+2) = 0.8
print('局部模块度：', R(comm1, test_G))
print('局部模块度：', R(comm2, test_G))
else:
virtual_social.append(comm)
matrix_r = [[0 for i in range(len(real_social))]
for i in range(len(virtual_social))]
for i in range(len(virtual_social)):
for j in range(len(real_social)):
merge_com = virtual_social[i] + real_social[j]
virtual_comm = virtual_social[i]
virtual_r = evaluation.R(virtual_comm, G)
merge_r = evaluation.R(merge_com, G)
data_r = merge_r - virtual_r
matrix_r[i][j] = data_r
print(matrix_r)
print(len(matrix_r))
for m in range(len(matrix_r)):
row_list = matrix_r[m]
row = row_list.index(max(row_list))
print(row_list)
print(row)
real_social[row] + virtual_social[m]
#real_social[matrix_r[m].index(max(matrix_r[m]))]+(virtual_social[m])
#print(matrix_r[m])
return real_social
football_network = './small_data/football-edges.txt'
G = cnn_socialNet_read_data.get_graph(football_network, split=',')
print(merge_community(social_list, G))
#print(len(merge_community(social_list, G)))
#print('nmi:',nmi_test.nmi())
bout = weight_variable([classnum])
# out = tf.matmul(dropf, Wout) + bout
out = tf.add(tf.matmul(dropf, Wout), bout)
return out
'''
1、构造图
2、将图的每条边转化为图片，根据训练好的模型识别每条边的标签
3、简化图：去掉识别为社区间的边，构成新的简化图
4、重复上面的1、2步骤，直到Q参数稳定
'''
# 1:构造图
file_network = './0814data/test_nodes.txt'
G = cnn_socialNet_read_data.get_graph(file_network)
# print(G.edges())
def get_data(g):
sess = tf.InteractiveSession()
edges = []
data = []
for u, v in g.edges():
edges.append((u, v))
for j in range(len(edges)):
matrix, row, clown = cnn_socialNet_deal_data.get_jump1_3dimension_different_size_matrix(g, edges[j])
image = tf.convert_to_tensor(matrix)
image = tf.image.convert_image_dtype(image, tf.float32)
resize_image = tf.image.resize_images(image, [128, 128], method=3)
img_numpy = resize_image.eval(session=sess)
print(G.edges())
a = tf.Variable(3)
print(a)
# 测试获取标准社区数据
file_path = './0814data/1/community-standard.txt'
test_file_path = './0814data/test_community.txt'
social_list = cnn_socialNet_read_data.get_standard_network(file_path)
print(social_list)
# 测试获取网络
test_network = './0814data/test_nodes.txt'
network_file_path = '0814data/1/network.txt'
test_G = cnn_socialNet_read_data.get_graph(network_file_path)
print(test_G.edges())
# 测试为网络中每条边添加flag属性
edges = []
cnn_socialNet_read_data.add_flag_graph(test_G, social_list)
for (u, v, flag) in test_G.edges.data('flag'):
# print(u, v, flag)
edges.append(((u, v), int(flag)))
# 获取每条边的矩阵
train_x = []
for i in range(len(edges)):
matrix, row, clown = cnn_socialNet_deal_data.get_jump1_3dimension_different_size_matrix(test_G, edges[i][0])
train_x.append((matrix, edges[i][1], edges[i][0]))
# print("matrix", matrix)
