In this paper, the annual road freight volume in Shanxi Province from 1979 to 2025 is taken as the research object, and the time series analysis and prediction research are carried out. The purpose of the research is to reveal the dynamic evolution law of road freight volume in Shanxi Province and provide quantitative reference and scientific basis for local traffic planning, logistics layout and industry management decision-making. In this paper, ARIMA time series modeling method is adopted, and the original freight volume series is systematically analyzed and tested by R software. Firstly, the original sequence is tested for stationarity. Although the sequence after first-order difference processing can pass the stationarity test, the test results show that the first-order difference sequence belongs to white noise sequence and contains insufficient effective information. In order to solve the problems of lack of sequence information and ineffective modeling, the data are further processed by second-order difference. The results show that the sequence after second-order difference not only meets the requirements of stationarity, but also does not appear as white noise, so the sequence has significant statistical regularity and modeling value. On this basis, the order of the model is determined by autocorrelation and partial autocorrelation functions, and the parameter estimation, model test and optimization are completed, and the optimal ARIMA forecasting model suitable for highway freight volume in Shanxi Province is established. The empirical results show that the second-order difference sequence modeling effect is more reasonable and reliable, and the model fits well, which can effectively reflect the freight volume.

In this paper, the annual road freight volume in Shanxi Province from 1979 to 2025 is taken as the research object, and the time series analysis and prediction research are carried out. The purpose of the research is to reveal the dynamic evolution law of road freight volume in Shanxi Province and provide quantitative reference and scientific basis for local traffic planning, logistics layout and industry management decision-making. In this paper, ARIMA time series modeling method is adopted, and the original freight volume series is systematically analyzed and tested by R software. Firstly, the original sequence is tested for stationarity. Although the sequence after first-order difference processing can pass the stationarity test, the test results show that the first-order difference sequence belongs to white noise sequence and contains insufficient effective information. In order to solve the problems of lack of sequence information and ineffective modeling, the data are further processed by second-order difference. The results show that the sequence after second-order difference not only meets the requirements of stationarity, but also does not appear as white noise, so the sequence has significant statistical regularity and modeling value. On this basis, the order of the model is determined by autocorrelation and partial autocorrelation functions, and the parameter estimation, model test and optimization are completed, and the optimal ARIMA forecasting model suitable for highway freight volume in Shanxi Province is established. The empirical results show that the second-order difference sequence modeling effect is more reasonable and reliable, and the model fits well, which can effectively reflect the freight volume.