Re-establish forecast accuracy measurement for dep. Demand Comparing the obs. bookings with the forecast under identical RM-control Parametric O&D forecast model Forecasts at: O&D traffic flow; Dep. Date; Day to departure; Policy. Arbitrary fare structure. Estimation uses only observable data, (un-constrained not used). Analytic solution Parsimonious use of forecast parameters Robustness also for extreme sparseness Self-consistent Inventory irrelevant as an explanatory variable Excellent forecast quality on real airline data In particular for the sell-up parameter. PODS simulations Old rule of thumb (10% forecast accuracy translates to 1% revenue) is still valid for errors in demand volume. New rule of thumb for errors in sell-up estimation: 10% forecast accuracy translates to 2-8% revenue. MSE as a ROM model Unresolved. Topic for continued research.
展开▼
