Expenses estimated for the equipment based on the optimal parameters ε -SVR

摘要：

Support vector regression machine (SVR) can establish the expenses estimate model according to the principle of the structure minimum risk. It reaches better accuracy in the small sample estimated, but there are not a set of integrity theories for choosing the parameters of SVR. This paper adopts the methods of varied step to search the optimal parameters. It use the combination accuracy to evaluate the estimated effect, research the influence of the parameters combined form to the estimated accuracy. Then we can assurance the optimized combined form of the parameters. The example indicates the optimized combined form of the parameters can improved the expenses estimated accuracy.

关键词： expenses estimated ε -Support vector regression machine (ε- SVR) optimal parameter kenatfunction

作者: SUN Lin-kai JIN Jia-shan GENG Jun-bao

作者单位: School of Naval Architecture & Power Naval University of Engineering Wuhan, China 430033

会议类型: 国际会议

会议名称: The 2010 International Conference on Computer Application and System Modeling(2010计算机应用与系统建模国际会议 ICCASM 2010)

会议地点: 太原

会议语种:英文

页码: 456-458

在线出版日期: 2010-10-22（万方平台首次上网日期，不代表论文的发表时间）