|
@@ -1143,7 +1143,7 @@ class Prophet(object):
|
|
|
lower_p = 100 * (1.0 - self.interval_width) / 2
|
|
|
upper_p = 100 * (1.0 + self.interval_width) / 2
|
|
|
|
|
|
- X = seasonal_features.as_matrix()
|
|
|
+ X = seasonal_features.values
|
|
|
data = {}
|
|
|
for component in component_cols.columns:
|
|
|
beta_c = self.params['beta'] * component_cols[component].values
|
|
@@ -1258,10 +1258,8 @@ class Prophet(object):
|
|
|
trend = self.sample_predictive_trend(df, iteration)
|
|
|
|
|
|
beta = self.params['beta'][iteration]
|
|
|
- Xb_a = (
|
|
|
- np.matmul(seasonal_features.as_matrix(), beta * s_a) * self.y_scale
|
|
|
- )
|
|
|
- Xb_m = np.matmul(seasonal_features.as_matrix(), beta * s_m)
|
|
|
+ Xb_a = np.matmul(seasonal_features.values, beta * s_a) * self.y_scale
|
|
|
+ Xb_m = np.matmul(seasonal_features.values, beta * s_m)
|
|
|
|
|
|
sigma = self.params['sigma_obs'][iteration]
|
|
|
noise = np.random.normal(0, sigma, df.shape[0]) * self.y_scale
|