基于LS-SVM的间断性需求备件预测

我国近年引进国外装备时间不长,其备件的历史数据较少、需求具有间断性且具有大量零值,给预测工作带来了一定的困难。文章提出应用最小二乘支持向量机（Least Squares Support Vector Machines LS-SVM）这一新的机器学习方法来实现间断性需求备件的预测,建立了舰艇间断性需求备件的预测模型,并对某型舰艇备件进行预测和分析,结果表明：LS-SVM在间断性需求备件预测上表现出优秀的学习和预测能力。

Forecasting for Spare Parts with Intermittent Demand Based on LS-SVM

It is hard to complete the forecasting job due to the short time of equipment introduction,as well as less history data and intermittent demand with a great deal of zero value of the spare parts.Therefore,this paper introduced a new method named LS-SVM（Least Squares Support Vector Machines） to meet the intermittent demand of forecasting for spare parts.With the help of establishing the forecasting model of ships,including both the forecasting and analysing work on some kind of ships,this paper then concluded that it could achieve better learning and forecasting application with intermittent demand for spare parts.