A new anomalous travel demand prediction method combining Markov model and complex network model

Accurate prediction of travel demand is crucial for the development of intelligent transportation systems. However, we are still lacking methods to predict travel demand in anomalous traffic conditions. In this study, we develop a new travel demand prediction method by combining Markov model and complex network model. First, the anomalous mobility network is generated and the anomalous mobility index is measured to quantify the anomaly of travel demand. Next, the time series matrix of the anomalous mobility indices is generated and integrated in the Markov chain model to predict travel demand. The proposed travel demand prediction method is compared with four benchmark models. Results indicate that the integration of Markov model and complex network model considerably improves the prediction accuracy of travel demand in anomalous traffic conditions.