N-Day Average Volume Based Time-Series Analysis for Passenger Flow of Metro

Taking daily data of Shanghai metro passenger flow as research object, an index of `n-day´ average passenger flow volume is introduced to construct “time-series”,the change rate of daily volume against `7-day´ average was used for analyzing the characteristics of working day passenger flow.On this basis, the research constructs ARIMA forecast model for Daily Passenger Flow of Shanghai Metro is constructed based on `N-Day´ Average Volume. The `7-day´ average volumes were calculated by iterated prediction model and recursive prediction model to forecast daily passenger flow volume. In the calculation process, The `7-day´ average volumes were directly calculated by model, and actual daily volumes were indirectly calculated by model with returned value. And, actual daily volumes are multiplied superposition factor by analysis result.The relative error of recursive prediction model against is less than of iterated prediction model by empirical test .The forecast error is within 2% in working days.