def test_data_frame_to_numeric():
a = numpy.concatenate((
numpy.repeat(["large"], 10),
numpy.repeat(["small"], 5),
numpy.repeat(["tiny"], 13),
numpy.repeat(["medium"], 3)))
b = numpy.concatenate((
numpy.repeat(["yes"], 8),
numpy.repeat(["no"], 23)))
rnd = numpy.random.RandomState(0)
c = rnd.randn(len(a))
input_df = pandas.DataFrame({"a_category": a,
"a_binary": b,
"a_number": c.copy()})
a_num = numpy.concatenate((
numpy.repeat([0], 10),
numpy.repeat([2], 5),
numpy.repeat([3], 13),
numpy.repeat([1], 3))).astype(numpy.int64)
b_num = numpy.concatenate((
numpy.repeat([1], 8),
numpy.repeat([0], 23))).astype(numpy.int64)
expected = pandas.DataFrame({"a_category": a_num,
"a_binary": b_num,
"a_number": c.copy()})
actual = column.categorical_to_numeric(input_df)
tm.assert_frame_equal(actual, expected, check_exact=True)
def setUp(self):
x, self.y = load_whas500()
x = column.categorical_to_numeric(column.standardize(x,
with_std=False))
self.x = x.values
self.columns = x.columns.tolist()
def _make_whas500(with_mean=True, with_std=True, to_numeric=False):
x, y = load_whas500()
if with_mean:
x = standardize(x, with_std=with_std)
if to_numeric:
x = categorical_to_numeric(x)
names = ['(Intercept)'] + x.columns.tolist()
return DataSetWithNames(x=x.values, y=y, names=names, x_data_frame=x)
def test_bool_series(self):
input_series = pandas.Series([True, True, False, False, True, False, True], name="human",
index=["Alpha", "Beta", "Gamma", "Delta", "Eta", "Mu", "Zeta"])
expected = pandas.Series([1, 1, 0, 0, 1, 0, 1], name="human",
index=["Alpha", "Beta", "Gamma", "Delta", "Eta", "Mu", "Zeta"])
actual = column.categorical_to_numeric(input_series)
tm.assert_series_equal(actual, expected, check_exact=True)
def test_series(self):
input_series = pandas.Series(["a", "a", "b", "b", "b", "c"], name="Thr33",
index=["Alpha", "Beta", "Gamma", "Delta", "Eta", "Mu"])
expected = pandas.Series([0, 0, 1, 1, 1, 2], name="Thr33",
index=["Alpha", "Beta", "Gamma", "Delta", "Eta", "Mu"])
actual = column.categorical_to_numeric(input_series)
tm.assert_series_equal(actual, expected, check_exact=True)
def whas500_sparse_data():
x, y = load_whas500()
x_dense = categorical_to_numeric(x.select_dtypes(exclude=[numpy.float_]))
data = []
index_i = []
index_j = []
for j, (_, col) in enumerate(x_dense.iteritems()):
idx = numpy.flatnonzero(col.values)
data.extend([1] * len(idx))
index_i.extend(idx)
index_j.extend([j] * len(idx))
x_sparse = coo_matrix((data, (index_i, index_j)))
return SparseDataSet(x_dense=x_dense, x_sparse=x_sparse, y=y)
def setUp(self):
x, self.y = load_whas500()
self.x_dense = column.categorical_to_numeric(x.select_dtypes(exclude=[numpy.float_]))
data = []
index_i = []
index_j = []
for j, (_, col) in enumerate(self.x_dense.iteritems()):
idx = numpy.flatnonzero(col.values)
data.extend([1] * len(idx))
index_i.extend(idx)
index_j.extend([j] * len(idx))
self.x_sparse = coo_matrix((data, (index_i, index_j)))
assert_array_equal(self.x_dense.values, self.x_sparse.toarray())
def setUp(self):
x, self.y = load_whas500()
self.x = categorical_to_numeric(x)