def main():
targetFile = "../Dados/main-features2.csv"
trainFile = '../boa.csv'
trainFeatures = first_dataset(trainFile)
trainDataSet = load_dataset(trainFile, trainFeatures).drop('id', 1)
trainFeatures.pop(0)
trainFeatures.pop()
targetFeatures = first_dataset(targetFile)
targetDataSet = load_dataset(targetFile, targetFeatures)
print(trainFeatures)
print(targetFeatures)
createNeuralNetworkClassifier(trainDataSet, trainFeatures, targetDataSet)
def generateOutput():
data_file = '../boa.csv'
target_file = '../Dados/main-features2.csv'
name_of_features = first_dataset(data_file)
target_features = first_dataset(target_file)
dataSet = load_dataset(data_file, name_of_features)
targetDataSet = load_dataset(target_file, target_features)
dataSet = dataSet.drop('id', 1)
name_of_features.pop(0)
name_of_features.pop()
#target_features.pop(0)
create_target(dataSet, name_of_features, targetDataSet, target_features)
def main():
targetFile = "../Dados/main-features2.csv"
trainFile = '../boa.csv'
trainFeatures = first_dataset(trainFile)
trainDataSet = load_dataset(trainFile, trainFeatures).drop('id', 1)
trainFeatures.pop(0)
trainFeatures.pop()
targetFeatures = first_dataset(targetFile)
targetDataSet = load_dataset(targetFile, targetFeatures)
print(trainFeatures)
print(targetFeatures)
decisionTree = createTreeClassifier(trainDataSet, trainFeatures,
targetDataSet)
visualize_tree(decisionTree, trainFeatures)
def train():
file_path = "../Dados/master-features.csv"
data_file = '../boa.csv'
name_of_features = first_dataset(data_file)
dataSet = load_dataset(data_file, name_of_features)
dataSet = dataSet.drop('id', 1)
name_of_features.pop(0)
name_of_features.pop()
print(name_of_features)
def gridSearch(targetFile, targetFeatures):
data_file = '../boa.csv'
name_of_features = first_dataset(data_file)
dataSet = load_dataset(data_file, name_of_features)
dataSet = dataSet.drop('id', 1)
name_of_features.pop(0)
name_of_features.pop()
print(name_of_features)
print(len(name_of_features), len(targetFeatures))
# prepare a uniform distribution to sample for the alpha parameter
#param_grid = {'alpha': sp_rand()}
# create and fit a ridge regression model, testing random alpha values
model = MLPClassifier()
parameters = {
#'learning_rate': ["constant", "invscaling"],
'hidden_layer_sizes': [(100, 10), (100, 20), (100, 30), (100, 1),
(100, 5)],
'learning_rate': ['constant', 'invscaling'],
'learning_rate_init': [0.05, 0.01, 0.1],
#'alpha': [10.0 ** -np.arange(1, 7)],
'activation': ['identity', 'logistic', 'tanh', 'relu']
}
rsearch = model_selection.GridSearchCV(estimator=model,
param_grid=parameters,
n_jobs=-1,
cv=10)
rsearch.fit(dataSet[name_of_features], dataSet['target'])
print(rsearch)
# summarize the results of the random parameter search
print(rsearch.best_score_)
print(rsearch.best_estimator_.alpha)
if targetFile:
prediction = rsearch.predict(targetFile[targetFeatures])
cols = ['Predicted']
cenas = targetFile['id_target']
features = pd.DataFrame(prediction, columns=cols)
file_name = "output2.csv"
features.set_index(cenas, inplace=True)
features.to_csv(file_name, sep=',', encoding='utf-8')