论文部分内容阅读
在大多数预报问题中,一般都采用逐步回归方法解决因子选择问题,且用复相关系数的大小判断预报方程的优劣.事实上,不能单凭复相关系数的大小决定预报方程的取舍. 本文应用自由度调整复相关系数的方法对如何合理地选择预报方程进行了研究,并用此方法对长江口蓝点马鲛春汛渔期预报方程进行了最佳选择,得到了较好结果.回报检验结果误差最大的年度为2天,预报的结果也是令人满意的.
In most forecasting problems, stepwise regression is usually used to solve the problem of factor selection, and the predictive equation is judged by the magnitude of complex correlation coefficient. In fact, the choice of forecasting equation can not be decided by the complex correlation coefficient only. This paper studies how to choose the forecasting equation reasonably by using the method of adjusting the complex correlation coefficient with the degree of freedom, and uses this method to choose the best forecasting equation of the spring flood season in the blue spot, The year with the largest error is 2 days and the forecast result is also satisfactory.