手法を比較

	構造変化ARIMAモデル	自己回帰和分移動平均モデル (ARIMA Model)	構造的ブレークに対するChow検定
分野	計量経済学	計量経済学	計量経済学
系統	Regression model	Regression model	Regression model
提唱年	1989-1998	1970	1960
提唱者	Perron (1989); extended by Bai & Perron (1998)	George Box and Gwilym Jenkins	Gregory C. Chow
種類	Time series model with regime detection	Time series forecasting model	Test for structural break in regression coefficients
原典	Bai, J., & Perron, P. (1998). Estimating and testing linear models with multiple structural changes. Econometrica, 66(1), 47-78.	Box, G. E. P., & Jenkins, G. M. (1970). Time Series Analysis: Forecasting and Control. Holden-Day.	Chow, G. C. (1960). Tests of equality between sets of coefficients in two linear regressions. Econometrica, 28(3), 591–605.
別名	ARIMA with structural breaks, break-adjusted ARIMA, piecewise ARIMA, ARIMA with regime shifts	ARIMA, Box-Jenkins model, integrated ARMA, ARIMA(p,d,q)	Chow breakpoint test, structural break test, Chow yapısal kırılma testi
関連	3	6	2
概要	A structural break ARIMA model extends the standard ARIMA framework by explicitly identifying and accommodating one or more abrupt shifts in the level, trend, or dynamics of a time series. Rather than forcing a single set of ARIMA parameters across the entire sample, it fits separate ARIMA specifications for each regime defined by the detected break dates.	The ARIMA(p,d,q) model is the standard workhorse for univariate time series forecasting. It combines autoregressive terms (past values), differencing to induce stationarity, and moving average terms (past shocks) into a unified linear framework. Developed by Box and Jenkins (1970), it remains one of the most widely applied models in econometrics and applied statistics.	The Chow test, introduced by Gregory Chow in 1960, checks whether the coefficients of a linear regression are the same across two subsamples — that is, whether a structural break occurs at a known point such as a policy change, crisis, or regime shift. It compares the fit of a single pooled regression with the combined fit of two separate regressions; a large improvement from splitting indicates the relationship differs between the two periods or groups.