Short term traffic flow prediction based on on-line sequential extreme learning machine

Abstract

Traffic flow cannot be predicted solely based on historical data due to its high dynamics and sensitivity to emergency situations. In this paper, a real traffic data collected from 2011 to 2014 is used, and an adaptive prediction model based on a variant of Extreme Learning Machine (ELM), namely On-line Sequential ELM with forgetting mechanism, is built. The model has the capability of updating itself using incoming data, and adapts to the changes in real time. However, limitations, such as the requirements of large number of neurons and dataset size for initialization, are discovered in practice. To improve the applicability, another scheme involving sequential updating and network reconstruction is proposed. The experimental results show that, compared with the previous method, the proposed one has better performance in time while achieving the similar accuracy.