def plot(self, periods=None, figsize=(10,10), **plot_kwds):
"""Plot graphical display of FEVD
Parameters
----------
periods : int, default None
Defaults to number originally specified. Can be at most that number
"""
import matplotlib.pyplot as plt
k = self.neqs
periods = periods or self.periods
fig, axes = plt.subplots(nrows=k, figsize=figsize)
fig.suptitle('Forecast error variance decomposition (FEVD)')
colors = [str(c) for c in np.arange(k, dtype=float) / k]
ticks = np.arange(periods)
limits = self.decomp.cumsum(2)
for i in range(k):
ax = axes[i]
this_limits = limits[i].T
handles = []
for j in range(k):
lower = this_limits[j - 1] if j > 0 else 0
upper = this_limits[j]
handle = ax.bar(ticks, upper - lower, bottom=lower,
color=colors[j], label=self.names[j],
**plot_kwds)
handles.append(handle)
ax.set_title(self.names[i])
# just use the last axis to get handles for plotting
handles, labels = ax.get_legend_handles_labels()
fig.legend(handles, labels, loc='upper right')
plotting.adjust_subplots(right=0.85)
def plot_forecast(self, steps=1, figsize=(10, 10)):
"""
Plot h-step ahead forecasts against actual realizations of time
series. Note that forecasts are lined up with their respective
realizations.
Parameters
----------
steps : int
default 1
figsize : tuple[int, int]
default (10, 10)
"""
import matplotlib.pyplot as plt
fig, axes = plt.subplots(figsize=figsize, nrows=self.neqs,
sharex=True)
forc = self.forecast(steps=steps)
dates = forc.index
y_overlay = self.y.reindex(dates)
for i, col in enumerate(forc.columns):
ax = axes[i]
y_ts = y_overlay[col]
forc_ts = forc[col]
y_handle = ax.plot(dates, y_ts.values, 'k.', ms=2)
forc_handle = ax.plot(dates, forc_ts.values, 'k-')
lines = (y_handle[0], forc_handle[0])
labels = ('Y', 'Forecast')
fig.legend(lines, labels)
fig.autofmt_xdate()
fig.suptitle('Dynamic %d-step forecast' % steps)
# pretty things up a bit
plotting.adjust_subplots(bottom=0.15, left=0.10)
plt.draw_if_interactive()
def plot_forecast(self, steps=1, figsize=(10, 10)):
"""
Plot h-step ahead forecasts against actual realizations of time
series. Note that forecasts are lined up with their respective
realizations.
Parameters
----------
steps : int
default 1
figsize : tuple[int, int]
default (10, 10)
"""
import matplotlib.pyplot as plt
fig, axes = plt.subplots(figsize=figsize, nrows=self.neqs,
sharex=True)
forc = self.forecast(steps=steps)
dates = forc.index
y_overlay = self.y.reindex(dates)
for i, col in enumerate(forc.columns):
ax = axes[i]
y_ts = y_overlay[col]
forc_ts = forc[col]
y_handle = ax.plot(dates, y_ts.values, 'k.', ms=2)
forc_handle = ax.plot(dates, forc_ts.values, 'k-')
lines = (y_handle[0], forc_handle[0])
labels = ('Y', 'Forecast')
fig.legend(lines, labels)
fig.autofmt_xdate()
fig.suptitle('Dynamic %d-step forecast' % steps)
# pretty things up a bit
plotting.adjust_subplots(bottom=0.15, left=0.10)
plt.draw_if_interactive()