基于孤立点消除合理选择训练样本的物流预测

Forecasting of Logistics Demand Based on IFCM-RBFNN

为了提高物流需求的预测精度,提出一种基于消除孤立点合理选择训练样本的物流预测(IFCM-RBFNN)。首先采用密度方法识别和剔除原始物流需求数据中的孤立点,消除孤立点对聚类结果的不利影响;然后采用模糊均值聚类(FCM)算法对物流需求数据进行聚类,选择最优训练样本集;最后采用RBF神经网络建立物流需求预测模型,并采用仿真实验对其性能进行测试。仿真结果表明,相对于对比模型,IFCM-RBFNN提高了物流需求的预测精度,预测结果更加可靠。

In this paper, in order to improve the accuracy of the forecasting of logistics demand, we proposed the forecasting method based on IFCM-RBFNN, then introduced the working mechanism of the method and at the end, through a simulationcompared it with other logistics demand forecasting methods.