论文部分内容阅读
提出一种基于自回归移动差分模型修正神经网络误差的物流需求预测模型(ARIMA-BPNN)。该模型采用自回归移动差分模型对物流需求量进行建模与预测,捕捉物流需求量的线性变化趋势;采用BP神经网络对物流需求量非线性、随机变化规律进行预测,最后利用BP神经网络预测结果对自回归移动差分模型的预测误差进行修正,得到物流需求量的最终预测结果;采用仿真实验对模型的性能进行测试;结果表明,相对于其它预测模型,ARIMA-BPNN可以更加全面、准确地描述物流需求量复杂的变化规律,提高了物流需求量的预测精度。
A logistics demand forecasting model (ARIMA-BPNN) is proposed based on the autoregressive moving differential model to correct neural network errors. This model uses the autoregressive moving differential model to model and forecast the demand of logistics, and captures the linear trend of logistics demand. The BP neural network is used to predict the demand of logistics in a non-linear and random manner. Finally, BP neural network is used to predict Results The forecast error of the autoregressive moving differential model was modified to get the final forecast result of logistics demand. The simulation experiment was used to test the performance of the model. The results showed that compared with other forecast models, ARIMA-BPNN can be more comprehensive and accurate Describe the complex changes in demand for logistics and improve the forecasting accuracy of logistics demand.