Short-term traffic volume time series forecasting based on phase space reconstruction

摘要

A method for the short-term prediction of traffic volume time series owning chaos characteristics based on reconstruction of phase space is described. The principle of nearest neighbor equal distance method in phase space is introduced, and this approach is firstly applied to forecast a real traffic volume time series and obtain the forecasting result of the traffic volume, and the forecasting result is compared with the results obtained by neural network and gray mode with one rank & one variable (abbreviated as GM (1,1)). The experiments prove that the short-term traffic volume forecasting based on phase space reconstruction is valid and feasible.