def __init__(self, logging=True):
self.dataset = None
self.session = None
self.header = None
self.isSess = False
if logging:
self.log = open('./log/' + date() + '.log', 'a')
else:
self.log = None
def __init__(self, split=0.7, logging=True, random_state=None):
self.dataset = None
self.session = None
self.train_dataset = None
self.test_dataset = None
self.train_session = None
self.test_session = None
self.flows = None
self.train_flows = None
self.test_flows = None
self.split_ratio = split
self.train_size = None
self.seed = random_state
self.skip_datas = []
use_cores = multiprocessing.cpu_count() // 3 * 2
self.pclf = RandomForestClassifier(n_jobs=use_cores,
random_state=random_state)
self.sclf = RandomForestClassifier(n_jobs=use_cores,
random_state=random_state)
self.le = LabelEncoder()
self.pscaler = MinMaxScaler()
self.sscaler = MinMaxScaler()
self.spreds_train = None
self.spreds_test = None
self.sprobs_train = None
self.sprobs_train_all = None
self.sprobs_test = None
self.sprobs_test_all = None
self.ppreds_train = None
self.ppreds_test = None
self.pprobs_train = None
self.pprobs_train_all = None
self.pprobs_test = None
self.pprobs_test_all = None
self.pkt_train_ptime_mean = None
self.pkt_test_ptime_mean = None
if logging:
self.log = open('/tf/md0/thkim/log/' + date() + '.log', 'a')
else:
self.log = None
def __init__(self, h_threshold=1.0, random_state=None, verbose=True):
self.ppreds = None
self.pprobs = None
self.spreds = None
self.sprobs = None
self.flows = None
self.classes = None
self.y_true = None
self.h_threshold = h_threshold
self.l_threshold = None
self.init_th = None
self.step = None
self.isInit = False
self.delta = None
self.verbose = verbose
np.random.seed(random_state)
if verbose:
self.log = open('./log/' + date() + '.log', 'a')
else:
self.log = None
def __init__(self, split=0.7, clf='rf', logging=True, random_state=None):
self.train_dataset = None
self.test_dataset = None
self.split_ratio = split
self.train_size = None
self.seed = random_state
self.skip_datas = []
np.random.seed(random_state)
use_cores = multiprocessing.cpu_count() // 4 * 3
if clf == 'rf':
self.sclf = RandomForestClassifier(n_jobs=use_cores,
random_state=random_state)
elif clf == 'dt':
self.sclf = DecisionTreeClassifier(random_state=random_state)
elif clf == 'et':
self.sclf = ExtraTreeClassifier(random_state=random_state)
elif clf == 'adt':
self.sclf = AdaBoostClassifier(
base_estimator=DecisionTreeClassifier(
random_state=random_state),
random_state=random_state)
elif clf == 'arf':
self.sclf = AdaBoostClassifier(
base_estimator=RandomForestClassifier(
n_jobs=use_cores, random_state=random_state),
random_state=random_state)
elif clf == 'gbt':
self.sclf = GradientBoostingClassifier(random_state=random_state)
self.le = LabelEncoder()
self.scaler = MinMaxScaler()
self.spreds_train = None
self.spreds_test = None
if logging:
self.log = open('./log/' + date() + '.log', 'a')
else:
self.log = None
def log_format(*args):
strname = '[' + cktime.date() + ']'
for arg in args[:-1]:
strname = strname + '[' + arg + ']'
strname = strname + args[-1] + '.log'
return strname