A hybrid grey relational analysis and support vector machines approach for forecasting consumption of spare parts

Abstract

Aiming at the problem that the influence factors of spare parts consumption can't be considered properly, a combined method based on grey relational analysis and support vector machines (SVM) was proposed to forecast spare parts consumption. Firstly, grey relation grad between the influence factors and spare parts consumption was calculated by grey relational analysis and the selected main influence factors were taken as the input of SVM while the output was the consumption. Lastly, the test samples were input into the trained model for forecasting. The results show that, compared with GM(1,1) model and artificial neural network (ANN) model, the proposed model has better forecast accuracy and dynamic adaptability, which can provide some references for the spare parts management section.