蔬菜的供給彈性非常小，收成之後，不僅產量會決定售價的高低，同類蔬

菜之間的替代效果，對於價格變化也有很大的影響力。因此若能事先預測

同類蔬菜未來的價格變化，即可計劃各類蔬菜的生產量。在本篇論文中，

我們試著將時空數列應用在非空間系統的經濟領域上。以臺灣地區三種常

見的蔬菜為例，分別以時空數列的 STARMA 模式與單變量 ARIMA 時間數

列，利用蔬菜批發價格建立模式，並比較其短期預測效果。最後，就價格

變動與產銷策略之關係進行討論。

The supply elasticity of vegetables is so small. Once the

production has been known, it would reflect on the price as

soon as possible. And at the same time, the substitute effect

between the vegetables has also great influence on the change

of the price. However, if we could forecast the variation of

the vegetables price,then the production-marketing strategy

would be planned in advance. In this paper, we apply the

spatial time series analysis on the field of economic, which is

included in the non-spatial system. An investigate about the

price variation for three kinds of vegetables in Taiwan.And the

comparison of short-term forecasting performance for the STARMA

model and traditional ARIMA model are also made. Finally, we

discuss in detail about the relationship between the change of

vegetable price and production-marketing strategy.

Ali, M.M.(1979). Analysis of stationary spatial temporal processes: estimation
and prediction,Biometrika,66,513-518.
Bennett, R.J. (1979). Spatial time series,Pion Lin1ited,London.
Bennett, R.J.(1984). Advances in the analysis of spatial time series ,Spatial
Statistics and lvlodels,235-251.
Besag ,J. (1974). Spatial interaction and the statistical analysis of lattice
systen1s, Journal of the Royal Statistical Society)B, 36,192-225.
Box, G.E.P. and Jenkins, G.Iv1.(1976). Time series analysis forecasting and
control,2nd ed.,Rolden-Day,San-Francisco.
Cliff, A.D. and Ord, J.K.(1981). Spatial processes: models and applications,
Pion Limited,London.
Deutsch ,S.J. and Pfeifer, P.E.(1980a). A Three-Stage iterative procedure for
space-tilDe modeling , Techno'metrics,22 ,35-47.
Deutsch ,S.J. and Pfeifer, P.E. (1980b). Identification and interpretation of
first order space-tilDe ARIv1A models 1 Technometrics, 22,397-408.
Deutsch ,S.J. and Pfeifer, P.E.(1981). Space-Time ARIv1A modeling with
contemporaneously correlated innovations, Techno'metrics,23, 401-409.
Flahault ,A.,et al. (1988). IvIodelling the 1985 influenza epidemic in Frence ,
Statistics in Medicine,7,1147-1155.
Griffith, D.A., Raining, R.P. and Bennett, R.J. (1985). Estimating missing
values in space-time data series, Time Series Analysis: Theory and
Practice 6,273-296.
Haslett, J. and Raftery, A.E.(1989). Space-Tilne modelling with long - memory
dependence : assessing Ireland's wind power resource , Applied
Statistics,38,l-50.
NIann,H.B. and vVald,A.(1943). On the statistical treatment of linear stochastic
dfference equations, Econometrika, II, 173-270.
Pfeifer, P.E. and Bodily, S.E. (1990). A Test of space-tinle ARNIA modelling
and forcasting of hotel data, J o'Uxnal of Forecasting ,9, 255-272.
Raftery, A.E.,Haslett, J. and NIcColI, E.(1982). ,;Vind power: A space-time
process, Time Series Analysis: Theory and Practice 2, 191-202.
Rogers, A.(1974). Statistical analysis of spatial dispersion: the quadrat
method, Pion Limited,London.
Stoffer, D.S.(1986). Estimation and identification of space-time ARNIAX
models in the presence of n1issing data , Journal of the American Statistical
Association, 81,762-772.
Tjostheim, D.(1978). Statistical spatial series modelling, Adv. Appl. Frob.
,10, 130-154.
Wei, Willian1 W.S.(1990). Tims series analysis: univariate and multivariate
methods, Addison-Wesley.
Zeger, S.L.(1985). Exploring an ozone spatial time series in the frequency
dOlnain ,Jo'urnal of the American Statistical Associations, 80, 323-33l.