基于PCA和SVM的备件需求预测模型
首发时间:2013-11-28
摘要:针对目前大量零值的备件预测方法精度不高的问题本文提出了主成分分析和支持向量机回归相结合的方法,对需求进行预测。首先整理分析备件的历史消耗数据进行主成分分析,得到主成分数据序列,其次,根据主成分数据序列建立训练集训练支持向量机,得到主成分预测结果,转化为备件的需求数据作为支持向量机的输入,从而预测备件的需求数据。将算法运用于某潜艇的指挥系统从不常用备件的预测中,结果显示该模型能够较准确的预测备件的需求量。
关键词: 备件库存管理及优化 需求预测 支持向量机 主成分分析
For information in English, please click here
Principal component analysis and support vector machines-based spares forecasting model
Abstract:Aiming at the lack of history data of low consumption spare parts, a combined mathord based on principle component analysis and support vector machines was proposed to forecast spare parts consumption.First,principle component analysis is used to reduce the data dimension when multi-factors are analysis,the secleted principal components were taken as input of SVM while output was consumption.The model was used to forcast a low consumption spare of submarine command system,the result turned out that the model can predict the consumption with high accuracy.
Keywords: Spares management and optimization demand forecasting SVM PCA
基金:
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于PCA和SVM的备件需求预测模型
评论
全部评论0/1000