A NEW METHOD FOR NON-LINEAR CLASSIFY AND NON-LINEAR REGRESSION Ⅱ :APPLICATION OF SUPPORT VECTOR MACHINE TO WEATHER FORECAST

Feng Hanzhong; Chen Yongyi

Vol.15, NO.3, 2004

Article Contents

Article Navigation > Journal of Applied Meteorological Science > 2004 > 15(3): 355-365

Feng Hanzhong, Chen Yongyi. A new method for non-linear classify and non-linear regression II : Application of support vector machine to weather forecast. J Appl Meteor Sci, 2004, 15(3): 355-365.

Citation:

Feng Hanzhong, Chen Yongyi. A new method for non-linear classify and non-linear regression II : Application of support vector machine to weather forecast. J Appl Meteor Sci, 2004, 15(3): 355-365.

Feng Hanzhong, Chen Yongyi. A new method for non-linear classify and non-linear regression II : Application of support vector machine to weather forecast. J Appl Meteor Sci, 2004, 15(3): 355-365.

Citation:

Feng Hanzhong, Chen Yongyi. A new method for non-linear classify and non-linear regression II : Application of support vector machine to weather forecast. J Appl Meteor Sci, 2004, 15(3): 355-365.

PDF( 236 KB)

A NEW METHOD FOR NON-LINEAR CLASSIFY AND NON-LINEAR REGRESSION Ⅱ :APPLICATION OF SUPPORT VECTOR MACHINE TO WEATHER FORECAST

Feng Hanzhong¹,
Chen Yongyi²

1.
Chengdu Meteorological Center, Chengdu 610071
2.
Training Center of China Meteorological Administration, Beijing 100081

Received Date: 2003-01-25
Rev Recd Date: 2003-09-17
Publish Date: 2004-06-30

Abstract

Abstract

A novel weather forecast method using the support vector machine (SVM) is introduced. Both of SVM model of area rainfall categorical forecast of 15 mm excess and SVM model of single-station temperature regression in Sichuan basin are built upon ECMWF analysis fields of 500 hPa height, 850 hPa temperature, and sea level pressure from April to September through 1990—2000. Extensive experiments are performed with performances evaluated by the Threat Scores (TS) or Correlation Coefficient. Empirical results demonstrate much improved performance compared with those given by standard statistic analysis and forecast methods.
- Support vector machines (SVM);
- Pattern recognition;
- Regression estimation;
- Rainfall categorical forecast;
- Temperature forecast

FullText(HTML)

References(8)

Figures and Tables

图 1 (a)500 hPa预报因子选取图示, (b) 地面气压预报因子选取图示, (c)850 hPa温度预报因子选取图示

DownLoad: Full-Size Img PowerPoint

图 2 四川盆地面雨量区域划分图示

DownLoad: Full-Size Img PowerPoint

图 3 训练样本逐渐增多对应的SVM回归预报模型的预报检验图示

DownLoad: Full-Size Img PowerPoint

表 1 利用径向基函数核进行分类预报试验时参数r、b值的变化对SVM模型的影响

表 2 用SVM方法对四川盆地面雨量 (大于15 mm) 建立推理模型的试验结果

表 3 用SVM回归方法对四川盆地内单站逐日平均气温建立预报模型的试验结果

表 4 因子与预报对象之间相关与否对应的SVM回归模型对预报结果的影响评价 (单站气温)

表 5 四川盆地各区域面雨量大于15 mm的SVM分类预报模型 (在不同因子状况下) 的参数变化

表 6 四川盆地各区域面雨量大于15 mm的SVM分类预报模型 (在不同因子状况下) 的预报检验结果

References

[1]	叶笃正, 曾庆存, 郭裕福.当代气候研究.北京:气象出版社, 1991.164-177.
[2]	黄嘉佑, 谢庄.卡尔曼滤波在天气预报中的运用.气象, 1993, 19(4):3-7. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXX199304001.htm
[3]	陆如华, 徐传玉, 张玲.卡尔曼滤波在天气预报中的运用技术.数值预报产品释用公报, 1996, 5/6(双月刊):28-36.
[4]	胡江林, 张礼平.神经网络模型预报湖北汛期降水量的应用研究.气象学报, 2001, 59(6):776-782. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXB200106013.htm
[5]	张韧, 蒋国荣, 余志豪, 等.利用神经网络计算方法建立太平洋副高活动的预报模型.应用气象学报, 2000, 11(4):474-477. http://qikan.camscma.cn/jams/ch/reader/view_abstract.aspx?file_no=20000469&flag=1
[6]	熊秋芬等.神经网络方法在静止气象卫星多通道资料估算降水中的应用.气象, 2002, 28(9):17-21. http://www.cnki.com.cn/Article/CJFDTOTAL-QXXX200209004.htm
[7]	Vapnik V N.Statistical Learning Theory.John Wiley & Sons, Inc., New York, 1998.
[8]	fttp://www.kernel-machines.org/