摘要
利用支持向量机回归算法建立备件需求模型,对未来备件需求进行了预测,并结合实例将支持向量回归算法与传统的最小二乘拟合方法作比较。结果表明,支持向量回归算法在预测精度上具有明显的优势,该方法能够较好地适应样本数量较少、需求呈非线性特征的备件预测问题。
This paper applies support vector regression algorithm to establish spare parts demand model, to predict the future spare parts requirement. Examples will be combined with support vector regression algorithm with the traditional least-squares fitting method for comparison. Results show that support vector regression excels least square in forecasting accuracy. This method can better adapt to relatively small sample size, and nonlinear characteristics data.
出处
《物流科技》
2010年第4期67-69,共3页
Logistics Sci-Tech
关键词
支持向量机
备件
预测
support vector machine
spare parts
forecast
