Resumen
En el presente artículo se muestra un algoritmo iterativo para la estimación de parámetros de modelos ARMA en series temporales que tengan alguna observación ausente. Posteriormente, se efectúa la demostración de la convergencia de dicho algoritmo. Se presenta un ejemplo de estimación basado en la simulación de series temporales con un ordenador y se exponen las conclusiones llevadas a cabo por el autor.
Abstract
In this paper, an iterative algorithm to estimate para meters of ARMA models with missing observations is given. Furthermore, the convergence of the algorithm is proved. A example based in simulation of time series with a computer and the conclusions of this technique are also presented.
This is a preview of subscription content, log in via an institution to check access.
Bibliografia
ANDERSON, O. D. (1982):Time Series Analysis, North Holland.
BARHAM, S. Y. (1981): «The EM algorithm»,M. Sc. Diss., University of Wales.
BARHAM, S. Y., y DUNSTAN, F. D. J.(1982): «Missing values in Time Series», Time Series Analysis: Theory and Practice 2, Ed. O. Anderson, North Holland.
BOX, G. E. P., y JENKINS, G. M. (1970):Time Series Analysis, Forecasting and Control, Holden Day, San Francisco.
BOYLES, R. A. (1983): «On the convergence of the EM algorithm»,J. Roy Statist. Soc., B-45 No. 1, 47–50.
CRISTOBAL, A. (1985):Análisis de Series Temporales en ausencia de observaciones. Estimación de parámetros. Tesina de Licenciatura, Universidad de Zaragoza.
DAMSLETH, E. (1980): «Interpolating missing values in a Time Series»,Schand. J. Statist., 7, 33–39.
DEMPSTER, A. P.; LAIRD, N. M. y RUBIN, D. B. (1977): «Maximum likelihood estimation for incomplete data, via the EM algorithm»,J. Roy. Statist. Soc., B-39, 1–38.
DUNSMUIR W., y ROBINSON, P. M. (1981): «Parametric estimators for Time Series with missing observations»,Adv. Appl. Prob., 13, 129–146.
HARVEY A. C., y PHILLIPS G. D. A. (1979): «Maximum likelihood estimation of regresion models with autoregressive-moving average disturbances»,Biometrika, 66, 49–58.
JONES R. H. (1962): «Spectral analysis with regularly missed observations»,Ann. Math. Stat., 33, 455–461.
LJUNG (1982):Biometrika, 69, 265–268.
MARSHALL, R. J. (1980): «Autocorrelación estimation of Time Series with randomly missing observarions»,Biometrika, 67, 567–571.
ORCHARD T., y WOODBURY, M. A. (1972): «A missing information principle: Theory and applications»,Proceeding of the 6th. Berkeley Symposium on Mathematical Statistics and Probability, 1, 697–715.
PARZEN, E. (1963): «On Spectral Analysis with missing observations and amplitude modulation»,Sankhya, A-25, 383–392.
PRIESTLEY M. B. (1981):Spectral Analysis and Time Series, Academic Press, London.
ROBINSON, P. M. (1980): «Estimation and Forecasting for Time Series containing censored of missing observations».Time Series: Proceedings of the International Conference at Nottingham University, in March 1979. Ed. O. Anderson, North Holland, 167–171.
SAKAI, H.; SOEDA, T., y TOKUMARU, H. (1979): «On the relations between fitting autoregressions and periodograms with applications»,Ann. Statist, 7, 96–107.
SHAMAN, P., y TAN, S. B. (1982): «Them mean square error of one step ahead predictions for and autoregressive process when the coefficients are estimated with missing data»,Time Series Analysis: Theory and Practice, 1, 53–63. Ed. O. Anderson, Nort Holland.
TAN, S. B. (1979): «Maximum likelihood estimation in autoregressive processes with missing observations»,Ph. D. Thesis, University of Pittsburgh.
WU, C. F. J. (1983): «On the convergence properties of the EM algorithm»,Ann. Statist, 11, N. 1, 95–103.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Cristóbal, A.C. Un algoritmo iterativo para la estimacion de modelos arma con ausencia de observaciones. TDE 3, 141–156 (1988). https://doi.org/10.1007/BF02884330
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF02884330