节点文献
回归诊断在事件概率回归预报中的应用
The Application of Regression Diagnosis Method to Weather Event Probability Regression Prediction
【摘要】 针对事件概率回归模型的特点,用残差分析和统计量诊断的方法归纳了回归残差的非对称分布现象,揭示了这种现象是由高杠杆点所引起,探讨了概率回归模型的残差不合理性的统计天气预报意义,从而提出适用于概率预报问题的事件概率回归改进模型。分析指出,事件概率回归模型的不合理性并不是个别例子的特殊性所造咸,而是由该模型的数学特点所决定。概率回归改进模型要优于普通的概率回归模型。
【Abstract】 The regression diagnosis for the event probability regression madal was carried out. It isshowed that probability regression model have no rationality in the sense of statistical forecastingby the distribution of the probability residua and diagnosis amotunts. Thus an advancement ofevent probability regression model is proposed and the results show that it is superior to generalregression model by mathematical proving and calculation of the examples.
【关键词】 回归诊断;
残差分布;
高杠杆点事件概率;
【Key words】 regression diagnosis; event probability; distribution of residua; advancement of regression model;
【Key words】 regression diagnosis; event probability; distribution of residua; advancement of regression model;
- 【文献出处】 气象 ,METEOROLOGICAL MONTHLY , 编辑部邮箱 ,1994年07期
- 【分类号】P456.9
- 【被引频次】1
- 【下载频次】78