Research on vessel price interval Forecasting Model based on the rough insensitive loss function SVR

摘要：

Because the continually fluctuant price of vessel, the forecast of interval price is more dependable than precise data. A arithmetic of the rough εinsensitive loss function-based support vector machines is introduced, which is able to forecast the interval value. In order to improve the precision of forecast, parameters of the arithmetic is optimized by chaos traverse. On the way, fore-learning objective function is established, to replace the minimum training error object function which easily leads to over training. According to above, the forecasting model of the vessel interval price is founded. Experimental results show that this RSVR model can forecast the vessel interval price has a high precision, and the fore-learning objective can reduce the over-training problems.

作者: ZHANG Tianyi SUN Shcngxiang

作者单位: Department of Equipment Economics and Management, Naval Univ.of Engineering, Wuhan, P.R.China

会议类型: 国际会议

会议名称: 2011 International Conference on Product Innovation Management(第六届产品创新管理国际会议)

会议地点: 武汉

会议语种:英文

页码: 315-318

在线出版日期: 2011-07-16（万方平台首次上网日期，不代表论文的发表时间）