from spektral.datasets import Citation
import numpy as np
import pickle as pkl
import networkx as nx
import scipy.sparse as sp
from scipy.sparse.linalg.eigen.arpack import eigsh
import sys
import scipy.io as scio
from scipy.io import loadmat
from scipy.sparse import coo_matrix
import tensorflow as tf
defparse_index_file(filename):"""Parse index file."""
index =[]for line inopen(filename):
index.append(int(line.strip()))return index
defsample_mask(idx, l):"""Create mask."""
mask = np.zeros(l)
mask[idx]=1return np.array(mask, dtype=np.bool)defload_data(dataset_str):"""
Loads input data from gcn/data directory
ind.dataset_str.x => the feature vectors of the training instances as scipy.sparse.csr.csr_matrix object;
ind.dataset_str.tx => the feature vectors of the test instances as scipy.sparse.csr.csr_matrix object;
ind.dataset_str.allx => the feature vectors of both labeled and unlabeled training instances
(a superset of ind.dataset_str.x) as scipy.sparse.csr.csr_matrix object;
ind.dataset_str.y => the one-hot labels of the labeled training instances as numpy.ndarray object;
ind.dataset_str.ty => the one-hot labels of the test instances as numpy.ndarray object;
ind.dataset_str.ally => the labels for instances in ind.dataset_str.allx as numpy.ndarray object;
ind.dataset_str.graph => a dict in the format {index: [index_of_neighbor_nodes]} as collections.defaultdict
object;
ind.dataset_str.test.index => the indices of test instances in graph, for the inductive setting as list object.
All objects above must be saved using python pickle module.
:param dataset_str: Dataset name
:return: All data input files loaded (as well the training/test data).
"""
names =['x','y','tx','ty','allx','ally','graph']
objects =[]for i inrange(len(names)):withopen("/home/stone/.spektral/datasets/Citation/cora/ind.{}.{}".format(dataset_str, names[i]),'rb')as f:if sys.version_info >(3,0):
objects.append(pkl.load(f, encoding='latin1'))else:
objects.append(pkl.load(f))
x, y, tx, ty, allx, ally, graph =tuple(objects)
test_idx_reorder = parse_index_file("/home/stone/.spektral/datasets/Citation/cora/ind.{}.test.index".format(dataset_str))
test_idx_range = np.sort(test_idx_reorder)if dataset_str =='citeseer':# Fix citeseer dataset (there are some isolated nodes in the graph)# Find isolated nodes, add them as zero-vecs into the right position
test_idx_range_full =range(min(test_idx_reorder),max(test_idx_reorder)+1)
tx_extended = sp.lil_matrix((len(test_idx_range_full), x.shape[1]))
tx_extended[test_idx_range-min(test_idx_range),:]= tx
tx = tx_extended
ty_extended = np.zeros((len(test_idx_range_full), y.shape[1]))
ty_extended[test_idx_range-min(test_idx_range),:]= ty
ty = ty_extended
features = sp.vstack((allx, tx)).tolil()
features[test_idx_reorder,:]= features[test_idx_range,:]
adj = nx.adjacency_matrix(nx.from_dict_of_lists(graph))
labels = np.vstack((ally, ty))
labels[test_idx_reorder,:]= labels[test_idx_range,:]
idx_test = test_idx_range.tolist()
idx_train =range(len(y))
idx_val =range(len(y),len(y)+500)
train_mask = sample_mask(idx_train, labels.shape[0])
val_mask = sample_mask(idx_val, labels.shape[0])
test_mask = sample_mask(idx_test, labels.shape[0])
y_train = np.zeros(labels.shape)
y_val = np.zeros(labels.shape)
y_test = np.zeros(labels.shape)
y_train[train_mask,:]= labels[train_mask,:]
y_val[val_mask,:]= labels[val_mask,:]
y_test[test_mask,:]= labels[test_mask,:]return adj, features, labels, train_mask, val_mask, test_mask
defnormalize_adj(adj):"""Symmetrically normalize adjacency matrix."""
adj = sp.coo_matrix(adj)
rowsum = np.array(adj.sum(1))
d_inv_sqrt = np.power(rowsum,-0.5).flatten()
d_inv_sqrt[np.isinf(d_inv_sqrt)]=0.
d_mat_inv_sqrt = sp.diags(d_inv_sqrt)return adj.dot(d_mat_inv_sqrt).transpose().dot(d_mat_inv_sqrt).tocoo()defpreprocess_adj(adj):"""Preprocessing of adjacency matrix for simple GCN model and conversion to tuple representation."""
adj_normalized = normalize_adj(adj + sp.eye(adj.shape[0]))return sparse_to_tuple(adj_normalized)
ALL_W= scio.loadmat('ALL_W.mat')
ALL_X= scio.loadmat('ALL_X.mat')
ALL_Y = scio.loadmat('ALL_Y.mat')
A=ALL_W['ALL_W']
X=ALL_X['ALL_X']
y=ALL_Y['ALL_Y']
A=coo_matrix(A)
num=[198,190,192,188,186,182,196,191,193,191,181,192,184,181,187]
X= X / X.max(1, keepdims=True)
y=y.squeeze()
train_mask = np.zeros(len(y), dtype=np.bool)
val_mask = np.zeros(len(y), dtype=np.bool)
test_mask = np.zeros(len(y), dtype=np.bool)for i inrange(0,15,1):
temp_index=np.where(y==i)
index=temp_index[0]
length=len(index)
Desperation=0.10*length
Des=int(Desperation)
Desperation1=0.5*length
Des1=int(Desperation1)
np.random.shuffle(index)
index2 = index[num[i]::1]
index1=index[num[i]::1]
index = index[0:num[i]:1]# 训练集索引
train_mask[index]=True
val_mask[index1]=True
test_mask[index2]=True#A, X, y, train_mask, val_mask, test_mask = load_data('cora')
y= tf.one_hot(y,15,1,0)
N = A.shape[0]# A 为形状为 (N, N) 的网络邻接矩阵
F = X.shape[-1]# X 为形状为 (N, F) 的节点特征
n_classes = y.shape[-1]# y 表示形状为 (N, n_classes) 的标签print(n_classes)# print(A, X, y, train_mask, val_mask, test_mask,sep="\n--------------\n")from spektral.layers import GCNConv
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Input, Dropout
A = GCNConv.preprocess(A).astype('f4')# 预处理邻接矩阵print("\n-----------\n",A)
X_in = Input(shape=(F,))# X 节点特征
A_in = Input((N,), sparse=True)# 邻接矩阵
X_1 = GCNConv(128,'relu')([X_in, A_in])# 图卷积
X_1 = Dropout(0.5)(X_1)
X_2 = GCNConv(n_classes,'softmax')([X_1, A_in])# 分类
model = Model(inputs=[X_in, A_in], outputs=X_2)
model.compile(optimizer='adam',# 梯度下降
loss='categorical_crossentropy',# 分类的常用损失函数
weighted_metrics=['acc'])# 打印准确率
model.summary()# 显示当前网络的结构# Prepare data#X = X.toarray()
A = A.astype('f4')
validation_data =([X, A], y, val_mask)# Train model
model.fit([X, A], y,
epochs=200,
sample_weight=train_mask,
validation_data=validation_data,
batch_size=N,
shuffle=False)# Evaluate model
eval_results = model.evaluate([X, A],
y,
sample_weight=test_mask,
batch_size=N)print(eval_results)
第二次尝试
import numpy as np
import tensorflow as tf
from tensorflow.keras.callbacks import EarlyStopping
from tensorflow.keras.losses import CategoricalCrossentropy
from tensorflow.keras.optimizers import Adam
from spektral.data.loaders import SingleLoader
from spektral.datasets.citation import Citation
from spektral.layers import GCNConv
from spektral.models.gcn import GCN
from spektral.transforms import AdjToSpTensor, LayerPreprocess
SAVA_PATH ='./file/'
X=np.load(SAVA_PATH +'X.npy', allow_pickle=True)
y=np.load(SAVA_PATH +'Y.npy', allow_pickle=True)
A=np.load(SAVA_PATH +'A.npy', allow_pickle=True)
tr_mask=np.load(SAVA_PATH +'mask_tr.npy', allow_pickle=True)
val_mask=np.load(SAVA_PATH +'mask_val.npy', allow_pickle=True)
validation_data =([X, A], y, val_mask)
learning_rate =0.001
seed =0
epochs =200
patience =10
tf.random.set_seed(seed=seed)# make weight initialization reproducible# n_labels num of classes n_input_channels dimension of dataset
model = GCN(n_labels=15, n_input_channels=144)
model.compile(
optimizer=Adam(learning_rate),
loss=CategoricalCrossentropy(reduction="sum"),
weighted_metrics=["acc"],)
model.fit([X, A], y,
steps_per_epoch=1,
validation_data=validation_data,
validation_steps=1,
epochs=epochs,
callbacks=[EarlyStopping(patience=patience, restore_best_weights=True)],)
第三次尝试
from spektral.datasets import Citation
import numpy as np
import pickle as pkl
import networkx as nx
import scipy.sparse as sp
from scipy.sparse.linalg.eigen.arpack import eigsh
import sys
import scipy.io as scio
from scipy.io import loadmat
from scipy.sparse import coo_matrix
import tensorflow as tf
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Input, Dropout
from tensorflow.keras import backend as K
from spektral.layers import ops
from spektral.layers.convolutional.conv import Conv
from spektral.utils import gcn_filter
classGCNConv(Conv):def__init__(
self,
channels,
activation=None,
use_bias=True,
kernel_initializer="glorot_uniform",
bias_initializer="zeros",
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,**kwargs
):super().__init__(
activation=activation,
use_bias=use_bias,
kernel_initializer=kernel_initializer,
bias_initializer=bias_initializer,
kernel_regularizer=kernel_regularizer,
bias_regularizer=bias_regularizer,
activity_regularizer=activity_regularizer,
kernel_constraint=kernel_constraint,
bias_constraint=bias_constraint,**kwargs
)
self.channels = channels
defbuild(self, input_shape):assertlen(input_shape)>=2
input_dim = input_shape[0][-1]
self.kernel = self.add_weight(
shape=(input_dim, self.channels),
initializer=self.kernel_initializer,
name="kernel",
regularizer=self.kernel_regularizer,
constraint=self.kernel_constraint,)if self.use_bias:
self.bias = self.add_weight(
shape=(self.channels,),
initializer=self.bias_initializer,
name="bias",
regularizer=self.bias_regularizer,
constraint=self.bias_constraint,)
self.built =Truedefcall(self, inputs):
x, a = inputs
output = K.dot(x, self.kernel)
output = ops.modal_dot(a, output)if self.use_bias:
output = K.bias_add(output, self.bias)return output
@propertydefconfig(self):return{"channels": self.channels}
@staticmethoddefpreprocess(a):return gcn_filter(a)
SAVA_PATH ='./file/'
X=np.load(SAVA_PATH +'X.npy', allow_pickle=True)
y=np.load(SAVA_PATH +'Y.npy', allow_pickle=True)
A=np.load(SAVA_PATH +'A.npy', allow_pickle=True)
X= X / X.max(1, keepdims=True)
A=coo_matrix(A)
train_mask=np.load(SAVA_PATH +'mask_tr.npy', allow_pickle=True)
val_mask=np.load(SAVA_PATH +'mask_val.npy', allow_pickle=True)
test_mask=np.load(SAVA_PATH +'mask_test.npy', allow_pickle=True)
N = A.shape[0]# A 为形状为 (N, N) 的网络邻接矩阵
F = X.shape[-1]# X 为形状为 (N, F) 的节点特征
n_classes = y.shape[-1]# y 表示形状为 (N, n_classes) 的标签
A = GCNConv.preprocess(A)
X_in = Input(shape=(F,))# X 节点特征
A_in = Input((N,), sparse=True)# 邻接矩阵
X_1 = tf.keras.layers.BatchNormalization()(X_in)
X_1 = GCNConv(128,'relu')([X_in, A_in])# 图卷积
X_1= tf.keras.layers.BatchNormalization(axis=1,momentum=0.9)(X_1)
X_1=tf.keras.layers.Activation('relu')(X_1)
X_1= tf.keras.layers.BatchNormalization(axis=1,momentum=0.9)(X_1)
X_2 = GCNConv(n_classes,'softmax')([X_1, A_in])# 分类
X_2=tf.keras.layers.Activation('softmax')(X_2)
model = Model(inputs=[X_in, A_in], outputs=X_2)
adam=Adam(lr=0.001,decay=0.001)
model.compile(optimizer=adam,# 梯度下降
loss='categorical_crossentropy',# 分类的常用损失函数
weighted_metrics=['acc'])# 打印准确率
model.summary()# 显示当前网络的结构
A = A.astype('f4')
validation_data =([X, A], y, val_mask)# Train model
model.fit([X, A], y,
epochs=200,
sample_weight=train_mask,
validation_data=validation_data,
batch_size=N,
shuffle=False)# Evaluate model
eval_results = model.evaluate([X, A],y,
sample_weight=test_mask,
batch_size=N)print(eval_results)
from spektral.datasets import Citationimport numpy as npimport pickle as pklimport networkx as nximport scipy.sparse as spfrom scipy.sparse.linalg.eigen.arpack import eigshimport sysimport scipy.io as sciofrom scipy.io import loadmatfrom scipy.spa