论文部分内容阅读
介绍 Box-Jenkins 的实用预测模型,并提出用随机模拟法模型的参数进行估计。与通常使用的最小二乘法相比,模拟法无需考虑初始值的选取,就能保证迭代过程的收敛性。整个程序的编制使用模块化结构,使预测者无需了解太多的这方面知识,就能方便地使用。在案例的预测过程中,只需进行几次人工干预,就能输出多种分析结果。
Box-Jenkins practical forecasting model is introduced, and the stochastic simulation method is proposed to estimate the parameters. Compared with the commonly used least square method, the simulation method can ensure the convergence of the iterative process without considering the selection of the initial value. The entire program is structured using a modular structure so that forecaster can easily use without knowledge of too much of this knowledge. In the case of the forecast process, with only a few human intervention, you can output a variety of analytical results.