基于模糊认知图和LSTM混合方法的H公司物流需求预测研究

Research on Logistics Demand Forecasting of Company H Based on the Combination of Fuzzy Cognitive Map and Long Short-Term Memory

鉴于传统的预测方法难以有效满足需求预测的非线性关系以及预测的精确度,文章提出了一种FCM-LSTM组合模型预测方法,首先对影响物流需求的关键因素进行模糊状态求解,然后对求解得到的各分量以及物流公司的历史货运量展开LSTM神经网络迭代预测,最终结合实例构建评价模型以评估算法的预测效果。该预测方法探讨了经济指标、季节性模式和假日对物流需求的重大影响,强调了将认知与神经网络模型结合应用的潜力,为控制物流行业的复杂动态提供了参考意见。结果表明,该组合模型可以提高预测精准度,具有有效性和可行性,为解决物流需求预测提供了可行的参考方法。

It is difficult for traditional forecasting methods to effectively meet the nonlinear relationship and prediction accuracy of demand forecasting,so this paper proposes a FCM-LSTM combined model forecasting method.Firstly,the key factors affecting logistics d emand a re s olved b y t he f uzzy s tate m ethod.S econdly,e ach c omponent o btained b y s olution a nd t he h istorical f reight volume of logistics company are iteratively predicted by LSTM neural network.Finally,an evaluation model is constructed to evaluate the prediction effect of the algorithm.The study explores the significant impact of economic indicators,seasonal patterns,and holidays on logistics demand,highlights the potential for combining cognitive and neural network models and provides references for the control of the complex dynamics of the logistics industry.The results show that this combined model can improve the prediction accuracy,have effectiveness and feasibility,and provide a feasible reference method for solving the logistics demand forecasting.