Application of support vector machine to hydrometeorology
Jianzhou Chen · Hubei Water Power · 2008
Based on the 1951-2004 run-off records of Danjiang waterbasin,74 kinds atmospheric circulation and seawater temperature records quoted from State Climate Center,adopting the anomaly fitting rate of distant correlative choice factors corresponding to the forecast objectives one by one,the run-off forecast models of Danjiang waterbasin in different seasons are set up with the support vector machine;the cross verification by few samples with practical run-off records shows that,these models have fine computational stability,their operating rate is higher than the traditional mathod.