论文部分内容阅读
本文介绍一种新的预报值为区间的预测方法——《对分模糊相似判别法》。它采用“逐步回归”法选定的线性相关预报因子或普查的线性相关预报因子进行模糊相似判别,区间收缩,选出最优区间。使预测的平均可信度比“逐步回归”法建立预报方程预报有较大的提高。 用此方法编制的ALGOL60计算机程序应用于重庆、宜宾、达县、万县等四川中、东部地区的长期天气预报计算,取得了良好的效果,用此方法处理非线性相关问题,及极端值天气预报比较有效。
In this paper, a new forecasting method for forecasting interval is introduced, namely “halving fuzzy similarity discriminating method”. It uses the “stepwise regression” method to select the linear correlation forecasting factor or the census linear correlation forecasting factor to carry on the fuzzy similar discrimination, the interval shrinks, selects the best interval. The average confidence of the prediction is greatly improved than that predicted by the “stepwise regression” method to establish the forecasting equation. The ALGOL60 computer program developed by this method has been applied to the long-term weather forecast calculation in the middle and eastern areas of Chongqing, Yibin, Daxian, Wanxian and other regions in Sichuan Province and has achieved good results. This method is used to deal with the nonlinear related issues and extreme weather Forecast more effective.