计算机工程 ›› 2006, Vol. 32 ›› Issue (21): 31-32,9.doi: 10.3969/j.issn.1000-3428.2006.21.011

• 博士论文 • 上一篇    下一篇

一类基于SVM/RBF的气象模型预测系统

罗泽举1,宋丽红2,薛宇峰2,朱思铭1   

  1. (1. 中山大学数学与计算科学学院,广州 510275;2. 湛江海洋大学海滨校区,湛江 524005)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2006-11-05 发布日期:2006-11-05

One Kind of Weather Model Forecast System Based on SVM/RBF

LUO Zeju1, SONG Lihong2, XUE Yufeng2, ZHU Siming1   

  1. (1. School of Mathematics and Computer Science, Sun Yat-Sen University, Guangzhou 510275;2. Seashore Campus of Zhanjiang Ocean University, Zhanjiang 524005)
  • Received:1900-01-01 Revised:1900-01-01 Online:2006-11-05 Published:2006-11-05

摘要: 利用广东湛江地区近30年月平均气温的气象数据作为数据集,建立了一种新的基于径向基核函数的支持向量机模型预测系统。通过适当选择模型参数,其平均绝对百分比误差只有5.61%,在绝对误差温度小于等于2℃的条件下,预测的准确率达到了95%,显示出所建立的支持向量机模型预测系统的有效性。通过分析发现了湛江海岸地区气候和全球气候变暖一致的事实。

关键词: 径向基核函数, 支持向量机, 模型预测系统

Abstract: Using the weather data of average temperature of nearly thirty years recorded at Guangdong Zhanjiang seashore region as the data set, this paper sets up a new support vector machines models forecast system based on radial basis kernel function(RBF). By choosing the model parameter properly, the mean absolute percentage error is only about 5.61%, if the absolute error is less than 2℃, the rate of accuracy comes to 95%. This shows the validity of the SVMS forecast system. In addation, by analysing it finds that the climate gets warm gradually in Zhanjiang and this is consistent with the global climate.

Key words: Radial basis kernel function, Support vector machines, Model forecast system

中图分类号: