Figure 5: Plot of in-sample forecasts compared to training set data.