灰色关联支持向量机在备件库存消耗预测中的应用

备件库存消耗预测是多因素综合影响下的非线性、小样本预测问题,且不同备件消耗的影响因素有所差异。针对上述问题,提出了一种基于灰色关联分析和支持向量机回归相结合的备件库存消耗预测方法。首先利用灰色关联分析计算出各影响因子与备件库存消耗的灰色关联度,量化了各因子对备件库存消耗的影响程度;再将筛选出的主因子作为支持向量机的输入,并利用遗传算法对支持向量机参数进行寻优,避免人为选择参数的盲目性,从而有针对性地对机体不同备件进行预测。最后,通过实证分析,验证了该方法应用于备件库存消耗预测的有效性和优越性,预测精度高于传统的备件预测模型。

Application of Grey Correlation SVM in Reserve Consuming Prediction of Spare Parts

The reserve consuming prediction of spare parts is a multi-factor influenced,non-linear and small-sample problem,and different spare parts have different influencing factors.Aiming at the problem,we proposed a method for predicting the consumption of reserved spared parts based on grey correlation analysis and Support Vector Machine (SVM).Firstly,grey relational degree between the influencing factors and the spare part consumption was calculated by grey correlation analysis,and the effect of each factor on the consumption was quantized.Then,the main factors gained were taken as input of SVM,and genetic algorithm was used to optimize the parameters of SVM to avoid blindness in parameter selection.Thus could make right prediction for different spare parts.Lastly,the model was proved to be effective for reserve consuming prediction of spare parts,which is higher in accuracy than other traditional models.