The best combination of forecasting accuracy and computational efficiency is achieved by adjusting the sparse parameter υ.
This is determined by a combination of forecasting methods from the ATIS side, variability in travel times from the network side and travellers' experiences.
If a high-quality starting field model is available with only lower-quality updates, then the combination of forecasts from a core flow model and lower quality field models using an Ensemble Kalman filter can reduce the RMS errors of the resulting forecast.
It would also allow the exploration of gains in forecast accuracy by using a weighted combination of forecasts from several models and/or methods.
This has long been considered the way to go in the economic literature, with the premise that a combination of forecasts will outperform individual forecasts (Bates and Granger 1969).
A future line of research is to use the common trends in factor error correction models (FECM), and FAVAR with time-varying coefficients, among many other models, combining the predictions with different combinations of forecasting methods.
However, if the field model is relatively poor, then a combination of a forecast field model using a steady core flow model and a field model computed from ground-based data can reduce the overall misfit to the 'true' field.
Based on a representative EWS model, we find that, for a given degree of risk aversion, there is a unique combination of the forecast horizon and of the probability threshold that maximizes the policy maker's preferences, yielding the best possible model from a policy perspective.
We created several time series of AT for any combination of a forecast product and a lag period, for example, the time series of AT based on 1-day-ahead RDF consist of all ATs based on the RDF forecasts produced 1 day in advance; T t represents day of year (T = 1, 2… 365, spline function with df = 5) to account for seasonality; and η is the coefficient for the heat wave indicator variables.
