import machine_learning as ml
import preprocessing as pp
import bayesian as bn
if __name__ == '__main__':
print(Fore.GREEN +
f"--------------------------------------------------------")
print("BAYESIAN NETWORK INTERFACING TOOL - Thomas Tiotto (2019)")
print(f"--------------------------------------------------------")
print(Fore.RESET + "")
data_set_path = '../DBMedico/DBBCTI_20042014_VMMZ_GL.xls'
data_set_path = input(f"Data set [{data_set_path}] : ") or data_set_path
df = hp.read_dataset(data_set_path, "excel", sheet="DB3")
print(f"Number of records in data set before cleaning: {len( df )}")
# drop records according to specifications
df = pp.drop_records(df)
print(f"Number of records in data set after cleaning: {len( df )}")
print("")
# bin data in columns according to specifications
df = pp.bin_records(df)
NUM_VALUES = len(df.columns)
# make dictionaries mapping categorical codes to original values and viceversa
import datetime
from colorama import Fore
import helper as hp
import preprocessing as pp
import bayesian as bn
if __name__ == '__main__':
print( Fore.GREEN + f"-----------------------------------------" )
print( f"{datetime.datetime.now()}" )
print( f"-----------------------------------------" )
df = hp.read_dataset( '../DBMedico/DBBCTI_20042014_VMMZ_GL.xls', "excel", sheet="DB3" )
print( f"Number of lines before cleaning: {len( df )}" )
# drop records according to specifications
df = pp.drop_records( df )
print( f"Number of lines after cleaning: {len( df )}" )
# bin data in columns according to specifications
df = pp.bin_records( df )
NUM_VALUES = len( df.columns )
# make dictionaries mapping categorical codes to original values and viceversa
df_values, df_codes, code_to_value_map, value_to_code_map = pp.make_mappings( df, NUM_VALUES )
# slice original dataset into value and code datasets
elif args.dataset == 'cifar10':
dataset = dset.CIFAR10(root=args.dataroot, download=True,
transform=transforms.Compose([
transforms.Scale(args.imageSize),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),
]))
testset = dset.CIFAR10(root=args.dataroot, train=False, download=True,
transform=transforms.Compose([
transforms.Scale(args.imageSize),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),
]))
elif args.dataset == 'ecg':
trainx, trainy = helper.read_dataset(args.dataroot + "/training_data4.hdf5")
testx, testy = helper.read_dataset(args.dataroot + "/test_data4.hdf5")
trainx = torch.from_numpy(trainx[:,10:,30:-20,:])
# testx = torch.from_numpy(testx)
# testx = np.array(testx)
# test =[]
# for i, img in enumerate(testx):
# img = skimage.transform.resize(img,(110,110))
# test.append(img)
# trainy = torch.from_numpy(np.argmax(trainy, axis=1))
testy = torch.from_numpy(np.argmax(testy, axis=1))
testx = torch.from_numpy(testx[:,10:,30:-20,:])
dataset = torch.utils.data.TensorDataset(trainx, trainy)
print(testx.size())
testset = torch.utils.data.TensorDataset(testx, testy)
transforms.Scale(args.imageSize),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5),
(0.5, 0.5, 0.5)),
]))
testset = dset.CIFAR10(root=args.dataroot,
train=False,
download=True,
transform=transforms.Compose([
transforms.Scale(args.imageSize),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5),
(0.5, 0.5, 0.5)),
]))
elif args.dataset == 'ecg':
trainx, trainy = helper.read_dataset(args.dataroot + "/FILE.hdf5")
testx, testy = helper.read_dataset(args.dataroot + "/FILE.hdf5")
testx = torch.from_numpy(testx[:, 10:, 30:-20, :])
# train = []
# trainx = np.array(trainx)
# for i, img in enumerate(trainx):
# if i%1000 == 0:
# print(i)
# img = skimage.transform.resize(img,(110,110))
# train.append(img)
# # print(np.s)
trainx = torch.from_numpy(trainx[:, 10:, 30:-20, :])
trainy = torch.from_numpy(np.argmax(trainy, axis=1))
testy = torch.from_numpy(np.argmax(testy, axis=1))