Abstract
In this paper, we propose an approach for forecasting traffic flow. The proposed approach uses a double exponential smoothing (DES) model and a Markov forecasting model to integrate a short/long-term scheme. Smoothing parameters of the DES model is determined by using the GA and the gradient descent algorithm. Additionally, model ability is evaluated using two directional tests: the directional change accuracy (DCA) and the regression test. The prediction model is applied to forecasting the traffic flow of the Route 23 in Nagoya-shi, Japan.
Keywords - Traffic flow forecast, double exponential smoothing, Markov chain, GA/gradient descent algorithm, directional change accuracy