def __init__(self, df, y_name, x_name, z_name, w_name, **kwargs):
super(IVReg, self).__init__(df, y_name, x_name, **kwargs)
# Handle extra variable stuff for IV
self.z_name = force_list(z_name)
self.w_name = force_list(w_name)
self.sample_cols_labels += ('z_name', 'w_name')
self.sample_store_labels += ('z', 'w')
self.vars_in_reg += ('z', 'w')
self.add_constant_to = 'w'
def __init__(self, df, y_name, x_name, z_name, w_name, **kwargs):
super(IVReg, self).__init__(df, y_name, x_name, **kwargs)
# Handle extra variable stuff for IV
self.z_name = force_list(z_name)
self.w_name = force_list(w_name)
self.sample_cols_labels += ('z_name', 'w_name')
self.sample_store_labels += ('z', 'w')
self.vars_in_reg += ('z', 'w')
self.add_constant_to = 'w'
def __init__(self, df, y_name, x_name, **kwargs):
self.df = df
self.y_name = y_name
self.__dict__.update(kwargs)
self.sample_cols_labels = (
'y_name', 'x_name', 'fe_name', 'cluster', 'shac_x', 'shac_y',
'awt_name'
)
self.sample_store_labels = (
'y', 'x', 'A', 'cluster_id', 'shac_x', 'shac_y', 'AWT'
)
self.vars_in_reg = ('y', 'x')
self.add_constant_to = 'x'
# Set `vce_type`
self.vce_type = _set_vce_type(self.vce_type, self.cluster, self.shac)
# Unpack spatial HAC args
sp_args = unpack_shac_args(self.shac)
self.shac_x = sp_args[0]
self.shac_y = sp_args[1]
self.shac_band = sp_args[2]
self.shac_kern = sp_args[3]
# Force variable names to lists
self.x_name = force_list(x_name)
def Ftest(self, col_names, equal=False):
"""F test using regression results.
Args:
col_names (str or list): Regressor name(s) to test.
Keyword Args:
equal (bool): Defaults to False. If True, test if all coefficients
in ``col_names`` are equal. If False, test if ``col_names`` are
jointly significant.
Returns:
tuple: A tuple containing:
- **F** (float): F-stat.
- **pF** (float): p-score for ``F``.
"""
cols = force_list(col_names)
V = self.vce.loc[cols, cols]
q = len(cols)
beta = self.beta.loc[cols]
if equal:
q -= 1
R = np.zeros((q, q + 1))
for i in range(q):
R[i, i] = 1
R[i, i + 1] = -1
else:
R = np.eye(q)
r = np.zeros(q)
return f_test(V, R, beta, r, self.df_r)
def flag_sample(df, *args):
varlist = []
for var in args:
if var is not None:
varlist += force_list(var)
sample = df[varlist].notnull().all(axis=1)
return sample
def __init__(self, df, y_name, x_name, **kwargs):
self.df = df
self.y_name = y_name
self.__dict__.update(kwargs)
self.sample_cols_labels = ('y_name', 'x_name', 'fe_name', 'cluster',
'shac_x', 'shac_y', 'awt_name')
self.sample_store_labels = ('y', 'x', 'A', 'cluster_id', 'shac_x',
'shac_y', 'AWT')
self.vars_in_reg = ('y', 'x')
self.add_constant_to = 'x'
# Set `vce_type`
self.vce_type = _set_vce_type(self.vce_type, self.cluster, self.shac)
# Unpack spatial HAC args
sp_args = unpack_shac_args(self.shac)
self.shac_x = sp_args[0]
self.shac_y = sp_args[1]
self.shac_band = sp_args[2]
self.shac_kern = sp_args[3]
# Force variable names to lists
self.x_name = force_list(x_name)
def Ftest(self, col_names, equal=False):
"""F test using regression results.
Args:
col_names (str or list): Regressor name(s) to test.
Keyword Args:
equal (bool): Defaults to False. If True, test if all coefficients
in ``col_names`` are equal. If False, test if ``col_names`` are
jointly significant.
Returns:
tuple: A tuple containing:
- **F** (float): F-stat.
- **pF** (float): p-score for ``F``.
"""
cols = force_list(col_names)
V = self.vce.loc[cols, cols]
q = len(cols)
beta = self.beta.loc[cols]
if equal:
q -= 1
R = np.zeros((q, q+1))
for i in range(q):
R[i, i] = 1
R[i, i+1] = -1
else:
R = np.eye(q)
r = np.zeros(q)
return f_test(V, R, beta, r, self.df_r)
def flag_sample(df, *args):
varlist = []
for var in args:
if var is not None:
varlist += force_list(var)
sample = df[varlist].notnull().all(axis=1)
return sample
def Ftest(self, col_names, equal=False):
cols = force_list(col_names)
V = self.vce.loc[cols, cols]
q = len(cols)
beta = self.beta.loc[cols]
if equal:
q -= 1
R = np.zeros((q, q+1))
for i in range(q):
R[i, i] = 1
R[i, i+1] = -1
else:
R = np.eye(q)
r = np.zeros(q)
return f_stat(V, R, beta, r, self.df_r)
def test_list(self):
expected = [1, 2, 3]
result = force_list(expected)
assert expected == result
def test_series(self):
a_series = pd.Series(np.arange(3))
expected = a_series.tolist()
result = force_list(a_series)
assert_equal(expected, result)
def test_string(self):
a_string = 'abcd'
expected = [a_string]
result = force_list(a_string)
assert_equal(expected, result)
def test_tup(self):
a_tuple = (1, 2, 3)
expected = list(a_tuple)
result = force_list(a_tuple)
assert_equal(expected, result)
def test_array(self):
an_array = np.arange(3)
expected = an_array.tolist()
result = force_list(an_array)
assert_equal(expected, result)
def test_list(self):
expected = [1, 2, 3]
result = force_list(expected)
assert_equal(expected, result)
def test_int(self):
an_int = 10
expected = [an_int]
result = force_list(an_int)
assert_equal(expected, result)
def test_string(self):
a_string = 'abcd'
expected = [a_string]
result = force_list(a_string)
assert expected == result
def test_tup(self):
a_tuple = (1, 2, 3)
expected = list(a_tuple)
result = force_list(a_tuple)
assert expected == result
def test_int(self):
an_int = 10
expected = [an_int]
result = force_list(an_int)
assert expected == result
def test_list(self):
expected = [1, 2, 3]
result = force_list(expected)
assert expected == result
def test_tup(self):
a_tuple = (1, 2, 3)
expected = list(a_tuple)
result = force_list(a_tuple)
assert expected == result
def test_int(self):
an_int = 10
expected = [an_int]
result = force_list(an_int)
assert expected == result
def test_array(self):
an_array = np.arange(3)
expected = an_array.tolist()
result = force_list(an_array)
assert expected == result
def test_string(self):
a_string = 'abcd'
expected = [a_string]
result = force_list(a_string)
assert expected == result
def test_series(self):
a_series = pd.Series(np.arange(3))
expected = a_series.tolist()
result = force_list(a_series)
assert expected == result
