本文整理匯總了Python中pandas.core.frame.DataFrame.shift方法的典型用法代碼示例。如果您正苦於以下問題：Python DataFrame.shift方法的具體用法？Python DataFrame.shift怎麽用？Python DataFrame.shift使用的例子？那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在類pandas.core.frame.DataFrame的用法示例。
在下文中一共展示了DataFrame.shift方法的1個代碼示例，這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚，您的評價將有助於係統推薦出更棒的Python代碼示例。
示例1: VAR
# 需要導入模塊: from pandas.core.frame import DataFrame [as 別名]
# 或者: from pandas.core.frame.DataFrame import shift [as 別名]
#.........這裏部分代碼省略.........
result = []
for i in xrange(h):
sum = self._alpha.reshape(1, k)
for j in xrange(self._p):
beta = self._lag_betas[j]
idx = i - j
if idx > 0:
y = result[idx - 1]
else:
y = self._data_xs(idx - 1)
sum = sum + np.dot(beta, y.T).T
result.append(sum)
return np.array(result)
def _forecast_std_err_raw(self, h):
"""
Returns the standard error of the forecasts
at 1, 2, ..., n timesteps.
"""
return np.array([np.sqrt(np.diag(value))
for value in self._forecast_cov_raw(h)])
@cache_readonly
def _ic(self):
"""
Returns the Akaike/Bayesian information criteria.
"""
RSS = self._rss
k = self._p * (self._k * self._p + 1)
n = self._nobs * self._k
return {'aic': 2 * k + n * np.log(RSS / n),
'bic': n * np.log(RSS / n) + k * np.log(n)}
@cache_readonly
def _k(self):
return len(self._columns)
@cache_readonly
def _lag_betas(self):
"""
Returns list of B_i, where B_i represents the (k, k) matrix
with the j-th row containing the betas of regressing the j-th
column of self._data with self._data lagged i time steps.
First element is B_1, second element is B_2, etc.
"""
k = self._k
b = self._beta_raw
return [b[k * i : k * (i + 1)].T for i in xrange(self._p)]
@cache_readonly
def _lagged_data(self):
return dict([(i, self._data.shift(i))
for i in xrange(1, 1 + self._p)])
@cache_readonly
def _nobs(self):
return len(self._data) - self._p
def _psi(self, h):
"""
psi value used for calculating standard error.
Returns [psi_0, psi_1, ..., psi_(h - 1)]
"""
k = self._k
result = [np.eye(k)]
for i in xrange(1, h):
result.append(sum(
[np.dot(result[i - j], self._lag_betas[j - 1])
for j in xrange(1, 1 + i)
if j <= self._p]))
return result
@cache_readonly
def _resid_raw(self):
resid = np.array([self.ols_results[col]._resid_raw
for col in self._columns])
return resid
@cache_readonly
def _rss(self):
"""Returns the sum of the squares of the residuals."""
return (self._resid_raw ** 2).sum()
@cache_readonly
def _sigma(self):
"""Returns covariance of resids."""
k = self._k
n = self._nobs
resid = self._resid_raw
return np.dot(resid, resid.T) / (n - k)
def __repr__(self):
return self.summary