def calculate(cls):
cls.g = g = GroupSorted(cls.gind) # pylint: disable-msg=W0201
cls.alla = [
(lag, sw.lagged_groups(cls.x, lag,
g.groupidx)) # pylint: disable-msg=W0201
for lag in range(5)
]
def calculate(self):
self.g = g = GroupSorted(self.gind) # pylint: disable-msg=W0201
self.alla = [
(lag, sw.lagged_groups(self.x, lag,
g.groupidx)) # pylint: disable-msg=W0201
for lag in range(5)
]
def __init__(self, endog, exog, group, sigma_i=None):
self.group = GroupSorted(group)
self.n_groups = self.group.n_groups
#self.nobs_group = #list for unbalanced?
nobs_i = len(endog) / self.n_groups #endog might later not be an ndarray
#balanced only for now,
#which is a requirement anyway in this case (full cov)
#needs to change for parameterized sigma_i
#
if sigma_i is None:
sigma_i = np.eye(int(nobs_i))
self.cholsigmainv_i = np.linalg.cholesky(np.linalg.pinv(sigma_i)).T
#super is taking care of endog, exog and sigma
super(self.__class__, self).__init__(endog, exog, sigma=None)
def __init__(self, endog, exog, group):
self.endog = endog
self.exog = exog
self.group = GroupSorted(group)
self.n_groups = self.group.n_groups
def calculate(self):
self.g = g = GroupSorted(self.gind)
self.alla = [(lag, sw.lagged_groups(self.x, lag, g.groupidx))
for lag in range(5)]
