class UpdateModel(object):
def __init__(self, iteration):
self.iteration = iteration
self.models = {}
self.times = {}
def run(self):
models_conf = self.iteration.conf.models_conf
for k, conf in models_conf.items():
self.runModel(k, conf)
def runModel(self, kind, conf):
self.setDatasets(conf)
model = conf.model_class(conf, cv_monitoring=False)
model.trainTestValidation(self.datasets)
self.models[kind] = model
# Execution time monitoring
time = model.training_execution_time + model.testing_execution_time
self.times[kind] = time
return None
def setDatasets(self, conf):
al_datasets = self.iteration.datasets
self.datasets = ClassifierDatasets(None, conf.sample_weight)
self.datasets.setDatasets(al_datasets.getTrainInstances(conf),
al_datasets.getTestInstances())
self.datasets.setValidationInstances(al_datasets.validation_instances)
def runNaiveBayes(self):
naive_bayes_conf = self.createNaiveBayesConf()
# Update training data - the naive Bayes classifier is trained on all the data
self.datasets.test_instances.annotations.setFamilies(
list(self.lr_predicted_labels))
all_datasets = ClassifierDatasets(None, naive_bayes_conf.sample_weight)
train_instances = copy.deepcopy(self.datasets.train_instances)
train_instances.union(self.datasets.test_instances)
all_datasets.setDatasets(train_instances, None)
self.evalClusteringPerf(all_datasets.train_instances)
# Train the naive Bayes detection model and predict
self.naive_bayes = GaussianNaiveBayes(naive_bayes_conf)
self.naive_bayes.training(all_datasets)
self.nb_time = self.naive_bayes.training_execution_time
num_test_instances = self.datasets.test_instances.numInstances()
self.datasets.test_instances.annotations.setFamilies(
[None] * num_test_instances)
if num_test_instances == 0:
self.nb_predicted_log_proba = []
else:
self.nb_predicted_log_proba = self.naive_bayes.pipeline.predict_log_proba(
self.datasets.test_instances.features.getValues())
start_time = time.time()
if num_test_instances == 0:
self.nb_predicted_labels = []
else:
self.nb_predicted_labels = self.naive_bayes.pipeline.predict(
self.datasets.test_instances.features.getValues())
self.nb_time += time.time() - start_time
self.nb_class_labels = self.naive_bayes.class_labels
def getMulticlassDatasets(self, model, alerts_ids):
datasets = ClassifierDatasets(None, model.conf.sample_weight)
train_instances = self.datasets.train_instances.getAnnotatedInstances(
label=labels_tools.MALICIOUS)
test_instances = self.datasets.test_instances.getInstancesFromIds(
alerts_ids)
datasets.setDatasets(train_instances, test_instances)
return datasets
def trainNaiveBayes(self):
naive_bayes_conf = self.getNaiveBayesConf()
datasets = ClassifierDatasets(None, naive_bayes_conf.sample_weight)
current_families = copy.deepcopy(
self.instances.annotations.getFamilies())
# families are altered
self.instances.annotations.setFamilies(self.assigned_categories)
datasets.setDatasets(self.instances, None)
naive_bayes = GaussianNaiveBayes(naive_bayes_conf)
naive_bayes.training(datasets)
# families are restored
self.instances.annotations.setFamilies(current_families)
return naive_bayes
def buildMulticlassClassifier(self):
multiclass_conf = self.getMulticlassConf()
datasets = self.iteration.datasets
predicted_instances = datasets.getInstancesFromIds(self.predicted_ids)
multiclass_datasets = ClassifierDatasets(None,
multiclass_conf.sample_weight)
multiclass_datasets.setDatasets(self.annotated_instances,
predicted_instances)
if self.multiclass_model is None:
self.multiclass_model = multiclass_conf.model_class(
multiclass_conf, cv_monitoring=True)
self.multiclass_model.training(multiclass_datasets)
self.multiclass_model.testing(multiclass_datasets)
if multiclass_datasets.validation_instances is not None:
self.multiclass_model.validation(multiclass_datasets)
return multiclass_datasets.train_instances, multiclass_datasets.test_instances
def generateDatasets(self):
instances = InstancesFromExperiment(self).getInstances()
test_instances = None
if self.conf.test_conf.method == 'dataset':
test_instances = InstancesFromExperiment(
self.test_exp).getInstances()
if self.conf.test_conf.method in [
'cv', 'temporal_cv', 'sliding_window'
]:
datasets = CvClassifierDatasets(self.conf.test_conf,
self.conf.families_supervision,
self.conf.sample_weight)
else:
datasets = ClassifierDatasets(self.conf.test_conf,
self.conf.sample_weight)
datasets.generateDatasets(instances, test_instances)
return datasets
def generateDatasets(self, classification_conf, instances, test_instances):
datasets = ClassifierDatasets(self,
classification_conf.sample_weight)
datasets.generateDatasets(instances, test_instances)
return datasets
def setDatasets(self, conf):
al_datasets = self.iteration.datasets
self.datasets = ClassifierDatasets(None, conf.sample_weight)
self.datasets.setDatasets(al_datasets.getTrainInstances(conf),
al_datasets.getTestInstances())
self.datasets.setValidationInstances(al_datasets.validation_instances)