Acta mathematica scientia,Series A ›› 2009, Vol. 29 ›› Issue (4): 1058-1064.

• Articles • Previous Articles     Next Articles

Strong Convergence Rates of the Maximum Quasi-Likelihood Estimator in Generalized Linear Models

 YIN Chang-Ming, LI Yong-Ming, WANG Peng-Yan   

  1. (School of Mathematics and Information |Science, Guangxi University, Guangxi Nanning 530004)
  • Received:2007-06-10 Revised:2008-09-19 Online:2009-08-25 Published:2009-08-25
  • Supported by:

    广西大学科研基金(DD051005)和广西自然科学基金(0832108, 0575051)资助

Abstract:

In a generalized linear model with q×1 responses, bounded and fixed p×q regressors Zi and general  link function, under  the moment
 condition on responses as weak as possible and other conditions, the authors prove that with probability one, the quasi-likelihood equation has a solution βn for all large sample size n. The rate of this solution tending to the true value is determined. In an important special case, this rate is the same as specified in the LIL for iid partial sums and thus cannot be improved anymore.  This result is an  improvement over the relevant results in literature.

Key words: Generalized linear model, Maximum quasi-likelihood estimate, Strong consistency,  , Strong convergence rate

CLC Number: 

  • 62F12
Trendmd