Another interpretation of the EM algorithm for mixture distributions. (English) Zbl 0585.62052

Summary: The EM algorithm for mixture problems can be interpreted as a method of coordinate descent on a particular objective function. This view of the iteration partially illuminates the relationship of EM to certain clustering techniques and explains global convergence properties of the algorithm without direct reference to an incomplete data framework.


62F10 Point estimation
65C99 Probabilistic methods, stochastic differential equations
Full Text: DOI


[1] Ball, G.H.; Hall, D.J., A clustering technique for summarizing multivariate data, Behavioral sciences, 12, 153-155, (1967)
[2] Bezdek, J.C., Pattern recognition with fuzzy objective function algorithms, (1981), Plenum Press New York · Zbl 0503.68069
[3] Dempster, A.P.; Laird, N.M.; Rubin, D.B., Maximum-likelihood from incomplete data via the EM algorithm, Journal of the royal statistical society, B39, 1-38, (1977) · Zbl 0364.62022
[4] Redner, R.A.; Walker, H.F., Mixture densities, maximum likelihood, and the EM algorithm, SIAM review, 26, 195-239, (1984) · Zbl 0536.62021
[5] Sclove, S.L., Application of the conditional population mixture model to image segmentation, IEEE transactions on pattern analysis and machine intelligence, 5, 428-433, (1983) · Zbl 0514.62065
[6] Titterington, D.M., Comments on “application of the conditional population mixture model to image segmentation”, IEEE transactions on pattern analysis and machine intelligence, 6, 656-657, (1984)
[7] Wolfe, J.H., Pattern clustering by multivariate mixture analysis, Multivariate behavioral research, 5, 329-340, (1970)
[8] Zangwill, W., Nonlinear programming: A unified approach, (1969), Prentice-Hall Englewood Cliffs, New Jersey · Zbl 0195.20804
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.