ARIMA乘法季节模型的R软件实现

Application and R Software Implementation of Multiplicative Seasonal ARIMA Model

摘要:
目的探讨ARIMA乘法季节模型的R软件实现方法，为模型的利用提供方法参考。
方法利用美国芝加哥市1987-2000年大气污染物臭氧（O₃）浓度数据建立ARIMA乘法季节模型，并进行预测，比较预测值和观察值的差异。
结果 ARIMA乘法季节模型在R软件中方便实现，模型预测值和观察值的平均相对误差为5.6%。
结论 R软件有相对丰富的软件包可以实现ARIMA乘法季节模型，使用者可以方便快捷地实现分析需求。

Abstract:
Objectives To explore the R software implementation of multiplicative seasonal ARIMA model and to provide a method for its utilization.
Methods The ARIMA model was established on account of the atmospheric pollutant ozone (O₃) concentration from 1987 to 2000 in Chicago, USA, and the difference between predicted value and observed value was compared.
Results ARIMA model was implemented conveniently in R software. and the average relative error between the predicted and observed values was 5.6%.
Conclusions R software had a relatively abundant useful software packages for fitting the multiplicative seasonal ARIMA model, and users could complete the analysis conveniently and quickly.