def predict_singleDN(self, model_folder, dname):
"""
对单个域名进行检测，输出检测结果及恶意概率
:param dname: 域名
:return:
"""
if not self.isload_:
self.load(model_folder)
dname = dname.strip('/').strip('.')
dname = dname.replace("http://", '')
dname = dname.replace("www.", "")
# dname = wash_tld(dname)
if dname == "":
label = 0
prob = 0.0000
p_value = 1.0000
print("\nxgboost dname:", dname)
# print("label:", label)
# print("mal_prob:", prob)
# print("p_value:", p_value)
print('label:{}, pro:{}, p_value:{}'.format(label, prob, p_value))
return label, prob, p_value
else:
feature = self.standardScaler.transform(
pd.DataFrame([get_feature(dname)]))
label = self.XGBoost_clf.predict(feature)
prob = self.XGBoost_clf.predict_proba(feature)
p_value = cal_pValue(self.train_score, prob[0][1], label[0])
print("\nxgboost sld:", dname)
# print("label:", label[0])
# print("mal_prob:", prob[0][1])
# print("p_value:", p_value)
print('label:{}, pro:{}, p_value:{}'.format(
label[0], prob[0][1], p_value))
return label[0], prob[0][1], p_value
def _load_image(img):
img = resize_img(img)
img = [get_feature(img)]
X = np.array(img)
X = X / 255
X = X.reshape(len(X), -1)
return X
def predict_singleDN(self, dname):
"""
对单个域名进行检测，输出检测结果及恶意概率
:param dname: 域名
:return:
"""
feature = self.standardScaler.transform(pd.DataFrame([get_feature(dname)]))
label = self.RF_clf.predict(feature)
prob = self.RF_clf.predict_proba(feature)
return label[0], prob[0][1]
def predict_singleDN(self, dname):
"""
对单个域名进行检测，输出检测结果及恶意概率
:param dname: 域名
:return:
"""
#self.SVM_clf = pickle.load(open(self.model_add, 'rb'))
#standardScaler = pickle.load(open(self.standard_scaler_add, 'rb'))
feature = self.standardScaler.transform(pd.DataFrame([get_feature(dname)]))
label = self.SVM_clf.predict(feature)
prob = self.SVM_clf.predict_proba(feature)
print("label:", label[0])
print("mal_prob:", prob[0][1])
return label[0], prob[0][1]
def load_dataset(dataset_lists, feature="edge_hist"):
data, labels, filenames = [], [], []
for dataset_list in dataset_lists:
for f in _read_dataset_list(dataset_list):
img = Image.open(f)
img = fe.get_feature(img, feature)
label = fe.get_label(f)
filename = fe.get_file_number(f)
data.append(img)
labels.append(label)
filenames.append(filename)
X, y, z = np.array(data), np.array(labels), np.array(filenames)
return _normalize_dataset(X, y, z)
