def plot_posterior_mean(add_mean=False):
# plot posterior mean
ax, x, y = visual_util.scaled_1d_kde_plot(ensemble_mean_arr,
shade=True,
color='red',
linewidth=0,
bandwidth=0.25,
density_scale=0.19)
return ax
def plot_posterior_mean(**kwargs):
# plot posterior mean
ensemble_mean_arr = np.concatenate(ensemble_mean_val)
ax, x, y = visual_util.scaled_1d_kde_plot(ensemble_mean_arr,
shade=True,
color='red',
linewidth=0,
bandwidth=0.25,
density_scale=0.19,
**kwargs)
return ax
def plot_posterior_sample(add_mean=False):
# plot uncalibrated posterior prediction
ax, x, y = visual_util.scaled_1d_kde_plot(ensemble_sample_arr,
shade=True,
color='grey',
linewidth=0,
density_scale=None)
if add_mean:
mean_height = y[np.argmin(abs(x - np.mean(ensemble_mean_arr)))]
visual_util.add_vertical_segment(np.mean(ensemble_mean_arr),
mean_height,
c='grey',
alpha=0.8)
return ax
def plot_posterior_sample(add_mode=False):
# plot uncalibrated posterior prediction
ensemble_sample_arr = np.concatenate(ensemble_sample_val)
ax, x, y = visual_util.scaled_1d_kde_plot(ensemble_sample_arr,
shade=True,
color='grey',
linewidth=0,
density_scale=None)
if add_mode:
median_val = x[np.argmax(y)]
median_height = y[np.argmin(abs(x - median_val))]
visual_util.add_vertical_segment(median_val,
median_height,
c='grey',
alpha=0.8)
return ax
def plot_true_density(add_mean=False):
# plot target
ax, x, y = visual_util.scaled_1d_kde_plot(y_valid,
shade=False,
linewidth=2,
linestyle='--',
color='grey',
density_scale=None)
if add_mean:
mean_height = y[np.argmin(abs(x - np.mean(y_valid)))]
visual_util.add_vertical_segment(np.mean(y_valid),
mean_height,
c='grey',
alpha=0.8,
linestyle='--')
# plot observed model predictions
for pred_val in np.concatenate(list(base_pred_dict.values())[1:]):
plt.plot([pred_val, pred_val], [0, 0.01], lw=3, c='black')