Short-term Traffic Flow Forecasting Using Transfer Ratio and Road Similarity

It is essential to make accurate traffic flow forecast or the Intelligent Transportation Systems (ITS). To improve the forecasting accuracy, most existing methods mainly focus on the time-series data. In this paper, we propose a novel model combining transfer ratio and road similarity (CoTRRS) to forecast the short-term traffic flow, which makes full use of the spatial information in the urban road networks. We first utilize the Continuous Bag-of-Words (CBOW) model to extract the road similarity and then combine it with transfer ratio to make accurate forecast using a back propagation neural network. To verify our model, we have performed extensive experiments on a real dataset, and the empirical study reveals that CoTRRS outperforms baselines.