摘要
利用2000年1月—2010年3月乌鲁木齐国际机场的观测资料,构建逐时能见度、温度、天气现象以及逐日最高温度、最低温度、降水量这六类预报对象的样本空间,使用SVM方法进行交叉验证和预报建模。结果表明建立的预测模型有较好的稳定性,并且对上述预报对象均有较好的预测效果。
Using the observational data from January 2000 to March 2010 at Urumqi International Airport, this paper constructed a sample space which include six forecasting objects： hourly visibility, temperature, and weather phenomena, and daily maximum temperature, minimum temperature, and precipitation, and employed the SVM method to do cross-validation and prediction modeling. The results showed that the prediction model kept good stability, and had better prediction effect for above objects.
出处
《沙漠与绿洲气象》
2011年第4期40-43,共4页
Desert and Oasis Meteorology
关键词
支持向量机(SVM)
分类预测
回归预测
温度
能见度
降水量
support vector machine （SVM）
classification forecast
regression forecast
temperature
visibility
precipitation
