近年來，非線性時間數列分析是一個快速發展的課題，其中最為人所矚目

的是門檻模式。從過去許多文獻得知，一個簡單門檻模式對於某些型態時

間數列的描述，如結構性改變的行為趨勢，比一般線性ARMA模式更能解釋

實際情況。在本篇論文中，我們將討論有關門檻模式及結構性改變分析的

問題。對於模式的建立，我們提出一個轉型期的觀念，替代傳統尋求一個

轉捩點的方法，進而提出一個結構性改變ARIMA模式有效建立的程序。最

後，我們以台灣出生率當作應用分析的範例，並且利用建立的結構性改變

ARIMA模式，及其他傳統門檻TAR模式，傳統線性分析方法等進行預測分析

及比較。

Non-linear time series analysis is a rapidly developing subject

in recent years. One of special families of non-linear models

is threshold model. Many literatures have shown that even

simple threshold model can describe certain types of time

series, such as structural change behavior, more faithful than

using linear ARMA models. In this paper, we discuss some

problems about the threshold model and structural change

analysis. Instead of finding the change point, we present the

change period concepts on the model- building. An efficient

algorithem on constructing the structure change ARIMA models is

proposed. Finally, we demonstrate an example about the birth

rate of Taiwan, and the comparison of forecasting performance

for the structure change ARIMA model with alternative models

are also made.

內政部人口政策委員會(1992) 人口資料彙集.
行政院主計處(1971-1992) 中華氏國統計月報.
蔡宏進(1990) 壹灣地區人口成長推計與未來人口政策推行之展望.
臺北:巨流圖書公司.
吳柏林、廖敏治(1991) 臺灣地區結婚率、出生率、人口成長率的時
問數列模式探討, 人口學刊, 第十四期, 頁109-132.
李美玲、王德睦、陳寬政(1993) 臺灣人口轉型後的人口政策與社會
政策, 人口與發展學街研討會論文集.

Box, G. E. P. and Tiao, G. C. (1975), Intervention analysis with applications to economic and environmental proble111S, J o'urnal of the A mer'ican Statistical Association, Vol. 70, No. 349, 70-79.
Broelneling, L. D. and Tsurumi, H. (1987), Eco'nometrics and structural change, marcel Dekker Inc.
Brown, R. L. Durbin, J. and Evans, .1. IvI. (1975), Techniques for testing the constancy of regression relationships over ti111e, Journal of the Royal
Statistical Society, Series B, Vol. 37, 149-197.
Carter, L. R. and Lee, R. D. (1986), Joint forecasting of U.S. rnartial fertility birth, and marriages using tinle series model, J o'urnal of the American
Statistical Association, Vol. Vol. 81,902-911.
Donovan,T.M. (1983),Short term forecasting : an introduction to the Box-Jenkins approach, John Wiley and Sons Ltd.
Davies N. Pemberton J. and Petruccelli J.D.(1988),An automatic procedure for identification, estimation and forecasting univariate self-exciting threshold autoregressive models, The Statistician,Vol.37, 199-204.
De Gooiler,J.G. and Kuldeep k. (1992), Some recent developments in nonlinear time series modeling,testing , and forecasting,International Journal of Forecasting , Vol.8,135-156.
Ferrer, A.F. and Hoyo,J.D. (1991), Analysis and prediction of the population in Spain:1910-2000, Journal of Forecasting, Vol. 10, 347-369.
Haines,L.M.Munoz,W.p. and Vav Gelderen,C.J.(1989), ARIMA modeling of birth data, Journal of Applied Statistics, Vol. 16, No.1, 55-67.
Hauser, J.A.(1992), Population,ecology and the new economics, guidelines for a steady-state economy, Futures,May,364-387.
Holbert,D. (1982),A Bayesian analysis of a switching linear model, Journal of Econometrics, Vol.19, 77-88.
Ploberger, W.Kramer,W. and Kontrus, K.(1989), A new test for structural stability in the linear regression model. Journal of Econometrics, Vol.40,307-318.
Priestley,M.B.(1988),Non-linear and non-stationary time series analysis, Academic press, Harcourt Brace Jovanovich, Publishers.
Petruccelli, J. D. (1990), A comparison of tests for SETAR-type non-linearity
in time series, Journal of Forecasting, Vol. 9, 25-36.
Roberts, G. R. (1988), Use of time-series methods in the preparation of quarterly
estimates of vital events, The Canadian J o'urnal of Statistics, Vol.
16, 75-82.
Shaban S. A. (1980), Change point problem and two-phase regression: an
annotated bibligraphy, International Statistical Review, Vol. 48, 83-93.
Tong, H. (1983), Threshold models in non-linear time series analysis, SpringerVerlag,
New York.
Tsay, R. S. (1986), Nonlinearity tests for time series, Biometrika, Vol. 73,
No.2. 461-466.
Tsay, R. S. (1988), Outliers, level shifts, and variance changes in tinle serise,
Jo'urnal of Forecasting, Vol. 7, 1-20.
Tsay, R. S. (1989), Testing and modeling threshold autoregressive proc(~sses)
Journal of the American Statistical Association, Vol. 84~ 231-240.
Watier L. and fuchardson S. (1991), A t.ime series construction of an alert
threshold with application to S. Bovismorbificans in France j Statistics
in Medicine, Vol. 10, 1493-1509.
Wei, W. W. S. (1990); Time series analysis) 'univariate and 'multi'variate
methods, Redwood City: Addison-Wesley.