Використання методів машинного навчання у прогнозуванні транспортної поведінки пасажирів

Abstract

Growth of urban population and its mobility, and a complexity of present transportation networks generates the travel behavior forecasting as an issue that needs to be addressed urgently. In case under consideration, the classical travel demand models (e.g. gravity, entropy and equilibrium models etc.) are inflexible because there is a necessity to deal with big data, diversity in preferences of public transport users, and constant changes taking place within the city environment.