def evaluateDynamic_changed_LinkPrediction(graph,
embedding,
rounds,
edges_add,
edges_rm,
n_sample_nodes=None,
no_python=False,
is_undirected=True,
sampling_scheme="u_rand"):
nodes = []
for e in edges_add[0]:
nodes.append(e[0])
nodes.append(e[1])
# for e in edges_rm[0]:
# nodes.append(e[0])
# nodes.append(e[1])
nodes = list(np.unique(nodes))
# pdb.set_trace()
test_digraph, node_l = graph_util.sample_graph(graph, len(nodes), nodes)
estimated_adj = embedding.predict_next_adj(node_l)
predicted_edge_list = evaluation_util.getEdgeListFromAdjMtx(
estimated_adj, is_undirected=is_undirected, edge_pairs=None)
MAP = metrics.computeMAP(predicted_edge_list, test_digraph)
prec_curv, _ = metrics.computePrecisionCurve(predicted_edge_list,
test_digraph)
return (MAP, prec_curv)
def evaluateDynamicLinkPrediction_TIMERS(graph,
embedding,
t,
rounds,
n_sample_nodes=None,
no_python=False,
is_undirected=True,
sampling_scheme="u_rand"):
node_l = None
if n_sample_nodes:
if sampling_scheme == "u_rand":
test_digraph, node_l = graph_util.sample_graph(
graph, n_sample_nodes)
else:
test_digraph, node_l = graph_util.sample_graph_rw_int(
graph, n_sample_nodes)
estimated_adj = embedding.predict_next_adj(t, node_l)
predicted_edge_list = evaluation_util.getEdgeListFromAdjMtx(
estimated_adj, is_undirected=is_undirected, edge_pairs=None)
MAP = metrics.computeMAP(predicted_edge_list, test_digraph)
prec_curv, _ = metrics.computePrecisionCurve(predicted_edge_list,
test_digraph)
return (MAP, prec_curv)
def evaluateStaticGraphReconstruction(digraph,
graph_embedding,
X_stat,
node_l=None,
sample_ratio_e=None,
file_suffix=None,
is_undirected=True,
is_weighted=False):
node_num = digraph.number_of_nodes()
# evaluation
if sample_ratio_e:
eval_edge_pairs = evaluation_util.getRandomEdgePairs(
node_num, sample_ratio_e, is_undirected)
else:
eval_edge_pairs = None
if file_suffix is None:
estimated_adj = graph_embedding.get_reconstructed_adj(X_stat, node_l)
else:
estimated_adj = graph_embedding.get_reconstructed_adj(
X_stat, node_l, file_suffix)
predicted_edge_list = evaluation_util.getEdgeListFromAdjMtx(
estimated_adj, is_undirected=is_undirected, edge_pairs=eval_edge_pairs)
MAP = metrics.computeMAP(predicted_edge_list, digraph)
prec_curv, _ = metrics.computePrecisionCurve(predicted_edge_list, digraph)
# If weighted, compute the error in reconstructed weights of observed edges
if is_weighted:
digraph_adj = nx.to_numpy_matrix(digraph)
estimated_adj[digraph_adj == 0] = 0
err = np.linalg.norm(digraph_adj - estimated_adj)
err_baseline = np.linalg.norm(digraph_adj)
else:
err = None
err_baseline = None
return (MAP, prec_curv, err, err_baseline)
def evaluateStaticGraphReconstruction(digraph,
graph_embedding,
X_stat,
node_l=None,
sample_ratio_e=None,
file_suffix=None,
is_undirected=True,
is_weighted=False):
"""Function to evaluate static graph reconstruction
Attributes:
digraph (Object): Networkx Graph Object
graph_embedding (object): Algorithm for learning graph embedding
X_stat (ndarray): Embedding values of the graph.
node_l (int): Total number of nodes.
sammple_ratio_e (float): SAmpling ration for testing. Only sample number of nodes are tested.
file_suffix (str): Suffix for file name.
is_undirected (bool): Flag to denote if the graph is directed.
is_weighted (bool): Flag denoting if the graph has weighted edge.
Returns:
ndarray: MAP, precision curve, error values and error baselines
"""
node_num = digraph.number_of_nodes()
# evaluation
if sample_ratio_e:
eval_edge_pairs = evaluation_util.getRandomEdgePairs(
node_num, sample_ratio_e, is_undirected)
else:
eval_edge_pairs = None
if file_suffix is None:
estimated_adj = graph_embedding.get_reconstructed_adj(X_stat, node_l)
else:
estimated_adj = graph_embedding.get_reconstructed_adj(
X_stat, node_l, file_suffix)
predicted_edge_list = evaluation_util.getEdgeListFromAdjMtx(
estimated_adj, is_undirected=is_undirected, edge_pairs=eval_edge_pairs)
MAP = metrics.computeMAP(predicted_edge_list, digraph)
prec_curv, _ = metrics.computePrecisionCurve(predicted_edge_list, digraph)
# If weighted, compute the error in reconstructed weights of observed edges
if is_weighted:
digraph_adj = nx.to_numpy_matrix(digraph)
estimated_adj[digraph_adj == 0] = 0
err = np.linalg.norm(digraph_adj - estimated_adj)
err_baseline = np.linalg.norm(digraph_adj)
else:
err = None
err_baseline = None
return (MAP, prec_curv, err, err_baseline)
def evaluateDynamic_changed_LinkPrediction(graph,
embedding,
rounds,
edges_add,
edges_rm,
n_sample_nodes=None,
no_python=False,
is_undirected=True,
sampling_scheme="u_rand"):
"""Function to evaluate dynamic changed link prediction
Attributes:
graph (Object): Networkx Graph Object
embedding (object): Algorithm for learning graph embedding.
edges_add (list): list of edges to be added.
edges_rm (list): list of edges to be removed.
n_sampled_nodes (int): List of sampled nodes.
train_ratio_init (float): sample to be used for training and testing.
rounds (int): Number of times to run the experiment
m_summ (str): summary to be used to save the result.
is_undirected (bool): Flag to denote if the graph is directed.
sampling_scheme(str): sampling scheme for selecting the nodes.
Returns:
ndarray: Mean Average precision
"""
nodes = []
for e in edges_add[0]:
nodes.append(e[0])
nodes.append(e[1])
# for e in edges_rm[0]:
# nodes.append(e[0])
# nodes.append(e[1])
nodes = list(np.unique(nodes))
# pdb.set_trace()
test_digraph, node_l = graph_util.sample_graph(graph, len(nodes), nodes)
estimated_adj = embedding.predict_next_adj(node_l)
predicted_edge_list = evaluation_util.getEdgeListFromAdjMtx(
estimated_adj, is_undirected=is_undirected, edge_pairs=None)
MAP = metrics.computeMAP(predicted_edge_list, test_digraph)
prec_curv, _ = metrics.computePrecisionCurve(predicted_edge_list,
test_digraph)
return (MAP, prec_curv)
def evaluateDynamicLinkPrediction(graph,
embedding,
rounds,
n_sample_nodes=None,
no_python=False,
is_undirected=True,
sampling_scheme="u_rand"):
node_l = None
if n_sample_nodes:
if sampling_scheme == "u_rand":
test_digraph, node_l = graph_util.sample_graph(
graph,
n_sample_nodes
)
else:
test_digraph, node_l = graph_util.sample_graph_rw_int(
graph,
n_sample_nodes
)
else:
test_digraph, node_l = graph_util.sample_graph(
graph,
n_sample_nodes)
estimated_adj = embedding.predict_next_adj(node_l)
print(len(estimated_adj), np.shape(estimated_adj))
predicted_edge_list = evaluation_util.getEdgeListFromAdjMtx(
estimated_adj,
is_undirected=is_undirected,
edge_pairs=None
)
print(len(predicted_edge_list), np.shape(predicted_edge_list), len(test_digraph.edges()),
np.shape(test_digraph.edges()))
# pdb.set_trace()
MAP = computeMAP(predicted_edge_list, test_digraph)
prec_curv, _ = computePrecisionCurve(
predicted_edge_list,
test_digraph
)
return (MAP, prec_curv)
def evaluateDynamicLinkPrediction(graph,
embedding,
rounds,
n_sample_nodes=None,
no_python=False,
is_undirected=True,
sampling_scheme="u_rand"):
"""Function to evaluate Dynamic Link Prediction
Attributes:
graph (Object): Networkx Graph Object
embedding (object): Algorithm for learning graph embedding
n_sample_nodes (list): sampled nodes
is_undirected (bool): Flag to denote if the graph is directed.
sampling_scheme (str): Sampling scheme to be used.
Returns:
ndarray: MAP, precision curve
"""
node_l = None
if n_sample_nodes:
if sampling_scheme == "u_rand":
test_digraph, node_l = graph_util.sample_graph(
graph, n_sample_nodes)
else:
test_digraph, node_l = graph_util.sample_graph_rw_int(
graph, n_sample_nodes)
estimated_adj = embedding.predict_next_adj(node_l)
print(len(estimated_adj), np.shape(estimated_adj))
predicted_edge_list = evaluation_util.getEdgeListFromAdjMtx(
estimated_adj, is_undirected=is_undirected, edge_pairs=None)
print(len(predicted_edge_list), np.shape(predicted_edge_list),
len(test_digraph.edges()), np.shape(test_digraph.edges()))
# pdb.set_trace()
MAP = metrics.computeMAP(predicted_edge_list, test_digraph)
prec_curv, _ = metrics.computePrecisionCurve(predicted_edge_list,
test_digraph)
return (MAP, prec_curv)
def main():
# data_list = ['cellphone', 'enron', 'fbmessages', 'HS11', 'HS12', 'primary', 'workplace']
data_list = ['bitcoin_alpha', 'bitcoin_otc', 'college_msg', 'enron_all', 'enron_all_shuffle']
funcs = ['AE', 'AERNN']
for data in data_list:
graphs = process('data/' + data)
length = len(graphs)
dim_emb = 128
lookback = 3
for func in funcs:
MAP_list = []
for i in range(length - lookback - 1):
if func == 'AERNN':
embedding = DynAERNN(d=dim_emb,
beta=5,
n_prev_graphs=lookback,
nu1=1e-6,
nu2=1e-6,
n_aeunits=[500, 300],
n_lstmunits=[500, dim_emb],
rho=0.3,
n_iter=250,
xeta=1e-3,
n_batch=100,
modelfile=None,
weightfile=None,
savefilesuffix=None)
elif func == 'RNN':
embedding = DynRNN(d=dim_emb,
beta=5,
n_prev_graphs=lookback,
nu1=1e-6,
nu2=1e-6,
n_enc_units=[500, 300],
n_dec_units=[500, 300],
rho=0.3,
n_iter=250,
xeta=1e-3,
n_batch=100,
modelfile=None,
weightfile=None,
savefilesuffix=None)
else:
embedding = DynAE(d=dim_emb,
beta=5,
n_prev_graphs=lookback,
nu1=1e-6,
nu2=1e-6,
n_units=[500, 300, ],
rho=0.3,
n_iter=250,
xeta=1e-4,
n_batch=100,
modelfile=None,
weightfile=None,
savefilesuffix=None)
embs = []
t1 = time()
# for temp_var in range(lookback + 1, length + 1):
emb, _ = embedding.learn_embeddings(graphs[i: i + lookback + 1])
embs.append(emb)
print(embedding._method_name + ':\n\tTraining time: %f' % (time() - t1))
pred_adj = graphify(embedding.predict_next_adj())
edge_index_pre = evaluation_util.getEdgeListFromAdjMtx(adj=pred_adj)
MAP = metrics.computeMAP(edge_index_pre, graphs[i + lookback + 1])
MAP_list.append(MAP)
print('第' + str(i) + '-' + str(i + lookback) + '个时间片的MAP值为' + str(MAP))
MAP_list.append(np.mean(MAP_list))
result = {'MAP值': MAP_list}
label = []
for i in range(len(MAP_list) - 1):
row = '第' + str(i) + '-' + str(i + lookback) + '个时间片'
label.append(row)
label.append('mean_MAP')
if not os.path.exists('result/' + data):
os.mkdir('result/' + data)
csv_path = 'result/' + data + '/' + str(func) + '.csv'
df = pd.DataFrame(result, index=label)
df.to_csv(csv_path)
def main():
graphs = process()
length = len(graphs)
dim_emb = 8
lookback = 3
MAP_list = []
for i in range(length - lookback - 1):
# embedding = DynAE(d=dim_emb,
# beta=5,
# n_prev_graphs=lookback,
# nu1=1e-6,
# nu2=1e-6,
# n_units=[500, 300, ],
# rho=0.3,
# n_iter=250,
# xeta=1e-4,
# n_batch=100,
# modelfile=['./intermediate/enc_model_dynAE.json',
# './intermediate/dec_model_dynAE.json'],
# weightfile=['./intermediate/enc_weights_dynAE.hdf5',
# './intermediate/dec_weights_dynAE.hdf5'],
# savefilesuffix="testing")
embedding = DynRNN(d=dim_emb,
beta=5,
n_prev_graphs=lookback,
nu1=1e-6,
nu2=1e-6,
n_enc_units=[500, 300],
n_dec_units=[500, 300],
rho=0.3,
n_iter=250,
xeta=1e-3,
n_batch=100,
modelfile=[
'./intermediate/enc_model_dynRNN.json',
'./intermediate/dec_model_dynRNN.json'
],
weightfile=[
'./intermediate/enc_weights_dynRNN.hdf5',
'./intermediate/dec_weights_dynRNN.hdf5'
],
savefilesuffix="testing")
embs = []
t1 = time()
# for temp_var in range(lookback + 1, length + 1):
emb, _ = embedding.learn_embeddings(graphs[i:i + lookback + 1])
embs.append(emb)
print(embedding._method_name + ':\n\tTraining time: %f' %
(time() - t1))
pred_adj = graphify(embedding.predict_next_adj())
edge_index_pre = evaluation_util.getEdgeListFromAdjMtx(adj=pred_adj)
MAP = metrics.computeMAP(edge_index_pre, graphs[i + lookback + 1])
MAP_list.append(MAP)
print('第' + str(i) + '-' + str(i + lookback) + '个时间片的MAP值为' + str(MAP))
with open('result/dynrnn_enron_MAP.txt', mode='w+') as file:
file.write('数据集共有' + str(length) + '个时间片\n')
file.write('lookback的值为' + str(lookback) + '\nMAP的值分别为：')
for MAP in MAP_list:
file.write(str(MAP) + ' ')
file.write('\n')
file.write('mean MAP: ' + str(np.mean(MAP_list)))
