def train(data, path):
if os.path.exists(path):
print 'Model path exists, do nothing.'
return
print "Loading features."
integerizer = tools.integerization.CIntegerization()
labels = []
samples = []
for sent in io.getsent(data):
for index in range(len(sent)):
f = feature.extractFeatures(sent, index, integerizer)
x = int(sent[index][2])
assert x == 0 or x == 1
if x == 0: x = -1
labels.append(x)
samples.append(f)
print "Training SVM."
problem = svm.svm_problem(labels, samples)
param = svm.svm_parameter()
param.svm_type = svm.C_SVC
param.kernel_type = svm.LINEAR #
param.C = 1
#param.degree=2
param.eps = 1.0
param.probability = 1
param.cache_size = 1000
param.shrinking = 0
model = svmutil.svm_train(problem, param)
print "Saving model."
os.mkdir(path)
svmutil.svm_save_model(os.path.join(path, "scr"), model)
integerizer.write(os.path.join(path, "int"))
def predict(data, path):
if not os.path.exists(path):
print 'Model path does not exist, do nothing.'
return
print "Loading features."
integerizer = tools.integerization.CIntegerization.read(os.path.join(path, 'int'))
model = svmutil.svm_load_model(os.path.join(path, "scr"))
print "Working."
for sent in io.getsent(data):
labels = []
for index in range(len(sent)):
f = feature.extractFeatures(sent, index, integerizer)
nodes, x = svmutil.gen_svm_nodearray(f)
probabilities = (ctypes.c_double*2)()
label = svm.libsvm.svm_predict_probability(model, nodes, probabilities)
x = 1
if probabilities[0] > cutoff:
x = 0
labels.append(str(x))
print ' '.join(labels)
def predict(data, path):
if not os.path.exists(path):
print 'Model path does not exist, do nothing.'
return
print "Loading features."
integerizer = tools.integerization.CIntegerization.read(
os.path.join(path, 'int'))
model = svmutil.svm_load_model(os.path.join(path, "scr"))
print "Working."
for sent in io.getsent(data):
labels = []
for index in range(len(sent)):
f = feature.extractFeatures(sent, index, integerizer)
nodes, x = svmutil.gen_svm_nodearray(f)
probabilities = (ctypes.c_double * 2)()
label = svm.libsvm.svm_predict_probability(model, nodes,
probabilities)
x = 1
if probabilities[0] > cutoff:
x = 0
labels.append(str(x))
print ' '.join(labels)
roles = np.argmax(g, axis=1)
print(g)
print(roles)
for i in range(len(roles)):
role = roles[i]
color.AddDat(i, roleToColor[role])
DrawGViz(graph, 0, imgName + '.png', 'Dot', True, color)
if __name__ == '__main__':
# read graph
fileName = "dataset/facebook_combined_small.txt"
#fileName = "dataset/Email-Enron.txt"
graph = LoadEdgeList(PUNGraph, fileName, 0, 1)
# feature extraction
v = feature.extractFeatures(graph)
# convert the node feature matrix to numpy matrix
v = toNumpyMatrix(v)
# pick number of roles with minimum error L = M + E
minError = sys.float_info.max
for r in range(2, 3):
g, f = factorization.nonNegativeFactorization(v, r)
error = computeDescriptionLength(v, g, f)
plotRole(graph, g, str(r) + '_roles')
if error < minError:
minError = error
finalG, finalF = g, f
numRoles = r
print 'using ' + str(numRoles) + ' roles'