US 12,248,897 B1
Deep learning-based demand forecasting system
Na Deng, Keller, TX (US); Benjamin Segal, Dallas, TX (US); Ou Bai, Euless, TX (US); Adam Thayer, Arlington, TX (US); and Venkata Pilla, Southlake, TX (US)
Assigned to AMERICAN AIRLINES, INC., Fort Worth, TX (US)
Filed by AMERICAN AIRLINES, INC., Fort Worth, TX (US)
Filed on Mar. 28, 2024, as Appl. No. 18/620,015.
Application 18/620,015 is a continuation of application No. 17/509,812, filed on Oct. 25, 2021, granted, now 11,948,111.
Claims priority of provisional application 63/107,143, filed on Oct. 29, 2020.
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 10/00 (2023.01); G06N 3/045 (2023.01); G06Q 10/04 (2023.01); G06Q 10/0631 (2023.01); G06Q 30/0201 (2023.01)
CPC G06Q 10/06315 (2013.01) [G06N 3/045 (2023.01); G06Q 10/04 (2013.01); G06Q 30/0206 (2013.01)] 12 Claims
OG exemplary drawing
 
7. A system for approximating a forecasting error of a passenger-demand forecasting model that comprises a convolutional neural network (“CNN”) model and a long short-term memory (“LSTM”) model, the system comprising a non-transitory computer readable medium having stored thereon a plurality of instructions, wherein the instructions are executed with one or more processors so that the following steps are executed:
calculating, using the passenger-demand forecasting model, a future passenger demand forecast for each key level in a set of key levels and for each departure date in a set of departure dates;
wherein each key level in the set of key levels is associated with a unique product for which demand is being forecasted; and
wherein each departure date of the set of departure dates is associated with a plurality of forecast periods;
wherein each forecast period of the plurality of forecast periods is associated with a period of time before departure;
wherein each departure date of the set of departure dates is associated with a plurality of fare products;
wherein each fare product of the plurality of fare products is a marketable combination of access rights associated with a charging moment; and
wherein calculating, using the passenger-demand forecasting model, the future passenger demand forecast for each key level in the set of key levels and for each departure date in the set of departure dates comprises:
applying a convolutional feature extractor on input data of the CNN model;
wherein the input data comprises historical fractional closures for each fare product of the plurality of fare products;
running the CNN model using the input data of the CNN model to create output(s) of the CNN model; and
running the LSTM model using input data of the LSTM model to create LSTM model output(s);
wherein input data of the LSTM model comprises:
 the output(s) of the CNN model; and
 features used to model seasonality;
wherein the output(s) of the LSTM model comprises predicted bookings with consideration of cancellation probabilities; and
wherein the future passenger demand forecast for each key level in the set of key levels and for each departure date in the set of departure dates comprises the predicted bookings with consideration of cancellation probabilities;
and
approximating, using a neural network trained on forecasting errors associated with the passenger-demand forecasting model, the forecasting error associated with the future passenger demand forecasts for the set of departure dates;
wherein the neural network has been trained on a training sample comprising forecast errors for historical demand forecasts to which a dropout model has been applied.