摘要
利用几十年气旋资料,应用经验统计及智能数据库技术,精选与气旋发生关系密切的因子,采用非线性回归变换,建立概率回归方程。将T016数值产品代入回归方程进行气旋概率预报。经实际应用并采用Brier评分,结果满意。
A statistics and intelligent database technique is applied to select the factorswhich should be closely related with cyclone by the use of several ten years data- The proba-' bility equation is set up with a non-linear regression change. The T106 numerical forecastproducts are put into the equation to do the cyclone forecasting. The result of Brier mark issatisfied after operational application.
出处
《气象科学》
CSCD
北大核心
1998年第4期403-408,共6页
Journal of the Meteorological Sciences