Highway traffic forecasting by support vector regression model with tabu search algorithms

Due to the nonlinearity and seasonality of highway traffic, many advanced forecasting models have been designed to predict traffic flow. Recently, the support vector regression model (SVR) has been successfully applied in some nonlinear time series forecasting problems. This study develops a support vector regression model with tabu search algorithms (SVRTS) to predict monthly highway traffics in Taiwan. The tabu search (TS) algorithm is employed to search the SVR parameters. Moreover, various tabu list sizes and neighbor solutions are used to examine the forecasting performance of SVRTS models. The experimental results revel that the SARTS model provides better forecasting results than the seasonal autoregressive integrated moving average (SARIMA) model. Thus, the developed SVRTS model is an appropriate alternative for forecasting highway traffic.