摘要
产品销售时序通常具有正态高斯分布、幅值较大、奇异点等混合噪音,为此,设计了一种鲁棒损失函数,得到一种新的支持向量机,即鲁棒ν-支持向量机。它可以有效地压制销售时序的多种噪音和奇异点,具有很强的鲁棒性,而且比标准ν-支持向量机具有更简洁的对偶优化问题。最后进行了汽车销售预测的实例分析,结果表明,基于鲁棒ν-支持向量机的预测模型是有效可行的。
To inhibit the normal Gaussian distributional noise, greater amplitude noise and singular points of product sales series, a robust loss function was designed, and a new support vector machine was obtained, named robust v - support vector machine （Rv-SVM）. Rv-SVM, which was with strong robustness and simpler dual optimization problem than standard v -SVM, various types of noises and singular points of product sales series could he inhibited effectively. Then, Rv -SVM was applied to predict ear sales, and the results showed that the prediction model based on Rv-SVM was effective and feasible.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2009年第6期1081-1087,共7页
Computer Integrated Manufacturing Systems
基金
国家863计划资助项目(2007AA04Z112)
国家自然科学基金资助项目(60574062
50875046)~~
关键词
支持向量机
预测模型
鲁棒损失函数
混合噪音
support vector machine
forecasting model
robust loss function
hybrid noises
