基于支持向量机的飞机备件消耗预测研究

Research on the Prediction of Aircraft Spare Parts Consumption Based on Support Vector Machine

针对影响飞机备件消耗的诸多因子难于在模型中体现的问题,采用支持向量机回归模型,应用于备件的消耗预测。该方法将影响备件消耗的主要因子作为支持向量机预测模型的输入因子,对应的备件消耗量作为输出因子,训练模型,然后输入测试样本进行预测。预测结果表明,相比于GM(1,1)模型和神经网络(ANN)模型,该模型具有较高的预测精度和动态适应性,可为相应的备件保障部门提供科学的决策依据。

In view of the problem that the consuming-related factors of aircrafts＇spare parts cant be revealed in the model, support vec- tor machine regression model is applied to predict the consumption of spare parts. In the model, the main factors that affect spare parts＂ consumption are used as the input factors of support vector machine prediction model, while the corresponding spare parts＇consumption as the output factors, and the test samples are input for prediction. The results show that, compared with GM （ 1,1 ） model and （ANN） model, this model has higher prediction accuracy and dynamic adaptability, which can provide references for spare parts management sections.