Aspects of Maximum Likelihood Estimation of the Logistic Growth Function
Logistic 增长函数最大似然估计的若干性质
The maximum likelihood estimators of the three-parameter logistic growth function with independent homoskedastic normally distributed error terms have, for sufficient observations, a variance-covariance matrix whose inverse is obtained. Some properties of the estimators are derived. An experimental study shows that small numbers of observations may lead to least squares estimates which do not have the convenient properties of estimates based on large samples.