def __init__(self,
units,
support=1,
activation=None,
use_bias=True,
adj=None,
kernel_initializer='glorot_uniform',
bias_initializer='zeros',
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
**kwargs):
if 'input_shape' not in kwargs and 'input_dim' in kwargs:
kwargs['input_shape'] = (kwargs.pop('input_dim'), )
super(GraphConvolution1, self).__init__(**kwargs)
self.adj1 = calculate_laplacian(adj)
self.adj = tf.sparse_tensor_to_dense(self.adj1)
self.units = units
self.activation = activations.get(activation)
self.use_bias = use_bias
self.kernel_initializer = initializers.get(kernel_initializer)
self.bias_initializer = initializers.get(bias_initializer)
self.kernel_regularizer = regularizers.get(kernel_regularizer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.activity_regularizer = regularizers.get(activity_regularizer)
self.kernel_constraint = constraints.get(kernel_constraint)
self.bias_constraint = constraints.get(bias_constraint)
self.supports_masking = True
self.support = support
def __init__(self, num_units, adj, num_nodes, input_size=None,
act=tf.nn.tanh, reuse=None):
super(tgcnCell, self).__init__(_reuse=reuse)
self._act = act
self._nodes = num_nodes
self._units = num_units
self._adj = []
self._adj.append(calculate_laplacian(adj))
def __init__(self, num_units, adj, inputs, output_dim, activation = tf.nn.tanh,
input_size = None, num_proj=None, reuse = None, **kwargs):
super(GCN, self).__init__(**kwargs)
if input_size is not None:
logging.warn("%s: The input_size parameter is deprecated.", self)
self._num_units = num_units
self._output_dim = output_dim
self._inputs = inputs
self._num_nodes = inputs.get_shape()[2].value
self._input_dim = inputs.get_shape()[1].value ###seq_len
self._batch_size = tf.shape(inputs)[0]
self._adj = []
self._adj.append(calculate_laplacian(adj))
self._activation = activation
self._gconv()
return rmse, mae, 1 - F_norm, r2, var
x_axe, batch_loss, batch_rmse, batch_pred = [], [], [], []
test_loss, test_rmse, test_mae, test_acc, test_r2, test_var, test_pred = [], [], [], [], [], [], []
# STGCN
n = num_nodes
n_his = seq_len
keep_rate = 1
total_W = np.array(adj_sampler.adj.todense()).astype(np.float32)
L = scaled_laplacian(total_W)
total_LK = cheb_poly_approx(L, Ks, n)
# TGCN or TGCN_att
TGCN_total_adj = calculate_laplacian(adj_sampler.adj.todense()).toarray()
for epoch in range(training_epoch):
if dropmode == 'node':
tmp_adj = adj_sampler.randomedge_sampler(ne_keep_prob)
elif dropmode == 'edge':
tmp_adj = adj_sampler.randomvertex_sampler(ne_keep_prob)
elif dropmode == 'out':
tmp_adj = adj_sampler.adj.todense()
W = np.array(tmp_adj).astype(np.float32)
L = scaled_laplacian(W)
# Alternative approximation method: 1st approx - first_approx(W, n).
Lk = cheb_poly_approx(L, Ks, n)
# tf.add_to_collection(name='graph_kernel', value=tf.cast(tf.constant(Lk), tf.float32))
def eval_model():
pass
if __name__ == "__main__":
args = parser()
scaler = MinMaxScaler(feature_range=(0, 1))
###### load data ######
train_data, test_data, adj = get_taxi_demand(file_name)
num_nodes = train_data.shape[1]
#### normalization
adj = calculate_laplacian(adj)
train_data = scaler.fit_transform(train_data)
test_data = scaler.fit_transform(test_data)
X_train, y_train1, y_train2, X_test, y_test1, y_test2 = preprocess_data(
train_data, test_data, args.lag)
print('trainX', X_train.shape)
print('trainY', y_train1.shape)
print('testX', X_test.shape)
print('y_test1', y_test1.shape)
print('y_test2', y_test2.shape)
start = time.clock()
y_predict = train_model(X_train, y_train2, X_test, y_test2)
data = pd.DataFrame(y_predict)
