非线性统计预报方法研究进展

The Evolution of Non-linear Statistical Forecast Methods

从线性统计分析向非线性统计分析发展,从而逐步建立非线性统计预报方法,已成为统计学方法发展的方向之一。虽然非线性统计预报这些年来有了较大发展,但却缺少系统的介绍。作者将自己的研究成果及应用实例归纳整理,对非线性统计预报中参数计算基本方法,一个自变量时的统计预报模型,拟线性、半线性和曲线性预报模式,分段逼近模式,插值、距离、相似预报方法等作了系统介绍。并给出了各种非线性统计预报模型在气象预报中的具体应用实例索引。

Developed from linear statistical analysis, non-linear statistical forecast methods become one aspect of statistical method. In recent years, non-linear statistical forecast method makes great progress but systematic introduction is still lack. In this paper, study results and application examples studied by author in recent years are summed up. The method of parameter calculating and the statistical forecast models of single independent variable are shown in this paper. Many forecast methods are introduced, such as fitting linear, half-linear and curve-linear forecast models, subsection approach model and analog forecast method. It also shows the example index of non-linear statistical forecast models.