本文選題：隨機(jī)抽取　切入點(diǎn)：AR模型　出處：《統(tǒng)計(jì)與決策》2016年24期

【摘要】：文章基于對(duì)平穩(wěn)時(shí)間序列數(shù)據(jù)的隨機(jī)抽取,選用AR模型研究其模型定階方法和參數(shù)評(píng)估準(zhǔn)則。根據(jù)數(shù)據(jù)有序性的特點(diǎn),提出利用交叉驗(yàn)證的方法確定自回歸模型階數(shù),并通過對(duì)原數(shù)據(jù)的無放回抽取實(shí)現(xiàn)對(duì)系數(shù)參數(shù)估計(jì)的評(píng)估。實(shí)例分析結(jié)果表明,交叉驗(yàn)證的定階與AIC準(zhǔn)則定階結(jié)果保持較高一致性,新的參數(shù)評(píng)估在一定的模型誤差范圍內(nèi)可以得到更為簡單有效的系數(shù)估計(jì)區(qū)間。
[Abstract]:Based on the random extraction of stationary time series data, AR model is selected to study the model order determination method and parameter evaluation criteria.According to the characteristics of data ordering, the method of cross validation is proposed to determine the order of autoregressive model, and the estimation of coefficient parameters is realized by extracting the original data without return.The results of example analysis show that the order determination of cross-validation is consistent with that of AIC criterion, and the new parameter evaluation can obtain a more simple and effective interval of coefficient estimation within a certain range of model errors.
【作者單位】： 太原理工大學(xué)數(shù)學(xué)學(xué)院;廈門大學(xué)管理學(xué)院;
【分類號(hào)】：O212.1;F224

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 張e，

本文編號(hào)：1700175