def runGridSearch():
attributes = Fingerprint.INFO_ATTRIBUTES + Fingerprint.HTTP_ATTRIBUTES + Fingerprint.JAVASCRIPT_ATTRIBUTES
min_numbers_fp = [2, 3, 4, 5]
max_counters = [80000, 100000, 130000, 170000, 200000, 220000]
thresholds = [100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 660]
activation_functions = ['identity', 'logistic', 'tanh', 'relu']
solvers = ["lbfgs", "adam"]
hidden_layers_sizes = [(3, 2), (4, 2), (5, 2), (6, 2), (3, 3), (4, 3), (5, 3), (6, 3), (2, 2, 2), (3, 3, 2), (4, 4, 2)]
algo_thresholds = [0.88, 0.89, 0.90, 0.91, 0.92, 0.93, 0.94, 0.95, 0.96, 0.97, 0.98]
model_name = "NeuralNetwork"
for min_number_fp in min_numbers_fp:
for max_counter in max_counters:
for i in range(0, len(thresholds)-1):
d = Data(attributes, computeSamples = True)
d.splitData(min_number_fp = min_number_fp, min_counter = 1000, max_counter = max_counter)
algo = Algorithm(attributes)
algo.setTrainSet(d.getTrainSample())
algo.setTestSet(d.getTestSample())
algo.computeRegressionInput(thresholds[i], thresholds[i+1])
for activation_function in activation_functions:
for solver in solvers:
for hidden_layer_size in hidden_layers_sizes:
algo.trainNNModel(hidden_layer_sizes=hidden_layer_size, activation=activation_function, solver=solver)
for algo_threshold in algo_thresholds:
algo.predictNN(threshold = algo_threshold)
algo.writeSubmission()
s_parameters = ("%d;%d;%d;%d;%s;%s;%s;%f" % (min_number_fp, max_counter, thresholds[i],\
thresholds[i+1], activation_function, solver, str(hidden_layer_size), algo_threshold))
print(s_parameters)
print(algo.evalPrecision(model_name, s_parameters))
def main(): d = Data(computeSamples = False) trainIndices, testIndices = d.splitData() algo = Algorithm(d.getTrainSample(), d.getTestSample()) # algo.computeRegressionInput() # algo.predictXGboost() algo.predictNN() # algo.predict() algo.writeSubmission() print(algo.evalPrecision())
def runSinglePrediction(attributes, min_number_fp, min_counter, max_counter, threshold_min, threshold_max, activation_function, solver,\
hidden_layers_size, algo_threshold, model_name):
d = Data(attributes, computeSamples=True)
d.splitData(min_number_fp=min_number_fp, min_counter=min_counter, max_counter=max_counter)
algo = Algorithm(attributes)
algo.setTrainSet(d.getTrainSample())
algo.setTestSet(d.getTestSample())
algo.computeRegressionInput(threshold_min, threshold_max)
algo.trainNNModel(hidden_layer_sizes=hidden_layers_size, activation=activation_function, solver=solver)
algo.predictNN(threshold=algo_threshold)
algo.writeSubmission()
s_parameters = ("%d;%d;%d;%d;%d;%s;%s;%s;%f" % (min_number_fp, min_counter, max_counter, threshold_min, \
threshold_max, activation_function, solver, str(hidden_layers_size),
algo_threshold))
print(s_parameters)
print(algo.evalPrecision(model_name, s_parameters))
