Forecasting of the Montreal Subway Smart Card Entry Logs with Event Data.

摘要: One of the major goals transport operators is to adapt supply scheduling passenger demand for existing networks during each specific period. Another problem mentioned by accurately estimating disposable ticket or pass availability demand. In this context, we propose generic data shaping, allowing use well-known regression models (basic, statistical and machine learning models) long-term forecasting with fine-grained temporal resolution. Specifically, paper investigates until one year ahead number passengers entering station a network quarter-hour aggregation taking planned events into account (e.g., concerts, shows, so forth). To compare quality prediction, real smart card event set from city Montr\'eal, Canada, that span three-year period two years training testing.