This article presents an SAS program for estimating a mixture of two normal linear regression models. The program is an EM algorithm and is written in the IML language in SAS. The usefulness of the algorithm is demonstrated by using an education example.
Get full access to this article
View all access options for this article.
References
1.
Beard, T. R. , Caudill, S. B., & Gropper, D. M. (1991). Finite mixture model estimation of multi-product cost functions. Review of Economics and Statistics, 75, 654-664.
2.
Caudill, S. B. , & Gropper, D. M. (1991). Test structure, human capital, and student performance on economics exams. The Journal of Economic Education, 22, 303-306.
3.
Dempster, A. P. , Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society,, 39, 1-38.
4.
Everitt, B. , & Hand, D. (1981). Finite mixture distributions. London: Chapman & Hall.
5.
Hardey, M. (1978). Comment (on estimating mixtures of normal distributions and switching regressions, by Quandt and Ramsey). Journal of the American Statistical Association, 73, 738-741.
6.
McLachlan, G. , & Basford, K. (1988). Mixture models: Inference and applications to clustering. New York: Marcel Dekker.
7.
Quandt, R. E. (1988). The econometrics of disequilibrium. New York: Basil Blackwell.
8.
Titterington, D. A. , Smith, A., & Makov, U. (1985). Statistical analysis of finite mixture distributions. New York: John Wiley.
