def prueba():
# create data_frame to input
data = DataFrame()
# create data_frame to output
output = DataFrame()
# create normalizer
normalizer = Normalizer()
data.load_data_set("breast-cancer-wisconsin-data.csv")
data.drop_columns_by_name(["id"])
output.data_set = data.cut_column('diagnosis')
data = data.data_set
data.data_set = normalizer.normalize_data(data)
data.join_data(output.data_set)
rn = NeuralNetwork()
kwargs = {"units": 2, "layers": 5, "activation": "sigmoid"}
rn.create_model(kwargs)
cv = KFoldCrossValidation(10, "diagnosis")
cv.k_fold_validation(data, model=rn)
cv.view_report()
def one_hot(self, data_set, column_tag):
df = DataFrame()
df.data_set = data_set
data_list = df.get_all_values().tolist()
dumies = self.dumies_tags(df, column_tag)
dumies = self.biuld_dumies(data_list, dumies)
return pd.DataFrame(dumies)
def one_hot2(self, data_set, column_tag):
df = DataFrame()
df.data_set = data_set
data_list = df.get_all_values().tolist()
dumies = self.dumies_tags_index(df, column_tag)
for i in range(len(data_list)):
data_list[i] = dumies[data_list[i][0]]
dic = {str(column_tag): data_list}
data = pd.DataFrame(dic)
return data
def prueba2():
# create data_frame to input
input = DataFrame()
# create data_frame to output
output = DataFrame()
# create normalizer
normalizer = Normalizer()
cv = CategoricalValues()
input.load_data_set("Churn_Modelling.csv")
# drop innecesary columns in the input
input.drop_columns_by_name(["RowNumber", "CustomerId", "Surname"])
output.data_set = input.cut_column('Exited')
# normalizer data
data = input.data_set
input.data_set = normalizer.normalize_data(data)
input.join_data(output.data_set)
cv = KFoldCrossValidation(10, "Exited")
cv.k_fold_validation(input)
cv.view_report()
def load_folds(self, data_frame, k):
from_index = 0
to_index = self.fold_size
for i in range(k + 1):
df = data_frame.sub_data_set(from_index, to_index)
fold = DataFrame()
fold.data_set = df
self.folds += [fold]
from_index += self.fold_size
to_index += self.fold_size
self.balance_last_fold(k)
help="Cantidad de neuronas por capa en la Red Neuronal")
net_group.add_argument(
"--funcion-activacion",
type=str,
default='relu',
metavar='',
help=
'Define la salida de los nodos en una Red Neuronal para un conjunto de entradas dado'
)
normalizer = Normalizer()
args = parser.parse_args()
output = DataFrame()
input = DataFrame()
input.load_data_set("breast-cancer-wisconsin-data.csv")
output.data_set = input.cut_column('diagnosis')
input.drop_columns_by_name(['id'])
data = input.data_set
input.data_set = normalizer.normalize_data(data)
input.join_data(output.data_set)
validation = KFoldCrossValidation(10, 'diagnosis')
if args.arbol:
print("Arbol has been chosen")
elif args.red_neuronal:
model = NeuralNetwork()
model.create_model(
kwargs={
"layers": args.numero_capas,
"units": args.unidades_por_capa,