|
|
|
|
|
|
| from scipy.stats import t
|
|
|
| class linear_regression():
|
| def __init__(self, fit_intercept = True):
|
| self.const = fit_intercept
|
|
|
| def fit(self, x, y):
|
| if self.const ==1:
|
| z = np.concatenate([np.ones((x.shape[0],1)), x], axis=1)
|
| self.index = x.columns.insert(0,'const')
|
| else:
|
| z = x
|
| self.index = x.columns
|
| self.phi = np.linalg.inv(z.T @ z) @ (z.T @ y)
|
| if self.const == 1:
|
| self.intercept_ = self.phi[0]
|
| self.coef_ = self.phi[1:]
|
| else:
|
| self.intercept_ = 'NA'
|
| self.coef_ = self.phi
|
| u = y - z @ self.phi
|
| RSS = np.sum(u**2)
|
| TSS = np.sum((y - np.mean(y))**2)
|
| self.R2 = 1 - RSS/TSS
|
| self.s2 = RSS/(z.shape[0] - z.shape[1])
|
| self.SE = np.sqrt(self.s2 * np.diagonal(np.linalg.inv(z.T @ z)))
|
| self.t = self.phi/self.SE
|
| self.p = (1 - t.cdf(abs(self.t), df = z.shape[0] - z.shape[1]))*2
|
|
|
| def predict(self, x):
|
| if self.const ==1:
|
| z = np.concatenate([np.ones((x.shape[0],1)), x], axis=1)
|
| else:
|
| z = x
|
| fcst = z @ self.phi
|
| return fcst.squeeze()
|
|
|
| def summary(self):
|
| """
|
| summaryの出力
|
| :return: summary dataframe
|
| """
|
| col_names = ["coef", "se", "t", "両側p値"]
|
| output = pd.DataFrame(np.c_[self.phi, self.SE, self.t, self.p],
|
| index=self.index,
|
| columns=col_names)
|
| return output |
