def process_wine(filename):
red_wine = numpy.genfromtxt(filename, delimiter=";", skip_header=True)
# deletes rows containing NaN values
red_wine = red_wine[~numpy.isnan(red_wine).any(axis=1)]
# classifies wines into 1 and 0
for row in red_wine:
if row[-1] > 5:
row[-1] = 1
else:
row[-1] = 0
# removes outliers from selected columns
list_del = {3, 4, 9}
for i in list_del:
li = Transformations.find_outliers(red_wine[:, i], 1.8)
for j in li:
red_wine = numpy.delete(red_wine, j, axis=0)
# adds features
# y = red_wine[:, -1]
# x = red_wine[:, 0:-1]
# x = Transformations.add_feature(x, [1, 4])
# red_wine = numpy.concatenate((x, y.reshape(y.shape[0], 1)), axis=1)
# selects features
# red_wine = Transformations.select_feature(red_wine, [8, 3])
# normalizes each column
for i in range(red_wine.shape[1]):
red_wine[:, i] = normalize(red_wine[:, i])
return red_wine
