Analysis of repeated measurement data is a recurrent focus of statisticians engaged in biological and biomedical applications, research have been well developed and well documented in the statistical literature. The early references include [l]-[5], which are based on the linear models or generalized linear models; on the other hand, [6]-[10] are those based on the nonlinear models which are more practical. But they are all restricted in the exponential family of distributions. That will not practice in most of the trials. Although being the most flexible among them, the estimation methods in [1] has many disadvantages for requiring the model to be generalized linear.Based on [1], this paper extended the method to the generalized nonlinear model. Assuming the response variables and covariate variables satisfied the generalized nonlinear model, we investigated the estimation and asymptotic properties of estimate of the regression parameters. Under some conditions, we proved the consistence and the asymptotically normality; also, we discussed the iterated methods of the estimation and attested it's convergence; at last, we testified the feasibility and validity of the estimation methods with simulation and some real examples analyses. |