A consideration on the maximum likelihood estimation in a mixture regression model using the EM algorithm

H. Hirose, Y. Komori

Far East Journal of Theoretical Statistics, Vol.7, No.2, pp.111-127 (2002.7)

To an extremely sensible problem of finding the maximum likelihood estimates in a specific mixture regression model such as the Weibull-power-law type regression model, the EM algorithm, as well as other optimization techniques, is applied to find the solution. Although some combination of several optimization techniques is found to be useful, the EM algorithm worked poorly.

Key Words
Newton-Raphson, simplex method, power-law, Weibull distribution,
Weibull-power-law, EM algorithm



Times Cited in Web of Science:

Cited in Books:

Cited in Proceedings: