Functional data analysis of daily curves in traffic: Transportation forecasting in the real-time

This study is motivated by an increasing focus on the intelligent transportation systems (ITS) that help people manage traffic flow and reduce travel times. Based on functional data analysis (FDA), we develop a prediction model to estimate the future traffic flow in the real-time. The proposed algorithm is built on existing methodology by applying the idea of initialization. Our algorithm is applied to the traffic flow from National Highway 5 in Taiwan. The experimental results show that the proposed method successfully reduces the prediction error by 30% or more.