Summary
The problem of predicting the number of change points in a piecewise linear model is studied from a Bayesian viewpoint. For a given a priori joint probability functionf R,C=fRf C/R, whereR is the number of change points andC=C′(R)=(C1,…,CR) is the change-point epoch vector, the marginal posterior probability functionf R.C/Y is obtained, and then used to find predictors forR andC(R).
Similar content being viewed by others
References
Bacon, D. W. andWatts, D. G. (1971). «Estimating the Transition between Two Intersecting Straight Lines», Biometrika 58, 525–43.
Broemeling, L. D. andChin Choy, J. H. (1981), «Detecting Structural Change in Linear Models», Commun. in Statistics A, vol. 10 # 24, 2551–61.
Ferreira, P. E., (1975). «A Bayesian Analysis of a Switching Regression Model: Known Number of Regimes», Journal of the American Statistical Association, vol. 70, 370–4.
Holbert, D. andBroemeling, L. D. (1977). «Bayesian Inferences Related to Shifting Sequences and Two-Phase Regression», Communications in Statistics A, vol. 6 # 3, 265–75.
McGee, V. E. andCarleton, W. T., (1970). «Piecewise Regression», Journal of the American Statistical Association, vol. 65 # 331, 1109–24.
Quandt, R. E., (1958). «The Estimation of the Parameters of a Linear Regression System Obeying Two Separate Regimenes», Journal of the American Statistical Association.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Aníbal Leiva, R. Bayesian prediction of the number of change points in the piecewise linear model. J. It. Statist. Soc. 3, 271–289 (1994). https://doi.org/10.1007/BF02589231
Issue Date:
DOI: https://doi.org/10.1007/BF02589231