| 傅轲钦,赵明.图解基于信息准则的ARMA模型选择[J].井冈山大学自然版,2017,(2):1-8 |
| 图解基于信息准则的ARMA模型选择 |
| GRAPHICAL INVESTIGATION OF ARMA MODEL SELECTION BASED ON INFORMATION CRITERION |
| 投稿时间：2016-09-29 修订日期：2016-11-21 |
| DOI：10.3969/j.issn.1674-8085.2017.02.001 |
| 中文关键词: 模型选择 ARMA 信息准则 EACF |
| 英文关键词: model selection ARMA information criterion EACF |
| 基金项目:贵州省科学技术基金计划项目（黔科合J字[2015]2064号） |
|
| 摘要点击次数: 2032 |
| 全文下载次数: 4254 |
| 中文摘要: |
| ARMA模型的选择是时间序列分析中的一个棘手问题，用于选择AR和MA模型最为常见的ACF和PCF函数对ARMA模型的选择不再有效。推广的自相关函数EACF在一定程度上给出了一种解决办法，但是这种将表格直观化的方法也存在一些不足，所以它并未被广泛接受和推广开来。ACF和PCF的图形化与EACF的表格直观化特性相结合是构建基于信息准则的ARMA模型选择的一种很好的方法，同时利用R软件的强大作图功能，可以实现图解ARMA模型的这一选择过程，最终可以得到一种既有效又直观的ARMA模型选择方法。 |
| 英文摘要: |
| The selection of ARMA model is a difficult problem in time series analysis. The selection of the most common ACF and PCF functions for AR and MA model is no longer valid for the selection of ARMA model. A certain solution extent is given for the auto-correlation function EACF. However, the method has also its disadvantages and not been widely accepted and popularized. Furthermore, the visual characteristics of EACF table based on the graphical features of ACF and PCF would be a good choice to select the ARMA model based on information criterion. Based on the powerful drawing function of R software, this selection process can be illustrated by ARMA model. Finally, an effective and intuitive method of ARMA model selection can be obtained. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
