On the statistical and biological behaviors of nonlinear mixed forest models

摘要: This study examined both statistical and biological behaviors of two nonlinear mixed models fitted by the first-order (FO) conditional expectation (FOCE) methods. The simpler FO method was found to alter mathematical forms base produce biologically unreasonable predictions for some subjects. not case more complex FOCE method. Since computations predicting random parameters making subject-specific (SS) were different methods, mixing them would potentially large prediction biases unexpected behaviors. demonstrated on data sets (basal area–age height–diameter sets). For basal data, accurate precise area produced as long a consistent used SS predictions, regardless model fitting. Otherwise, less produced, with totally against expectations. inconsistent methods similar statistics, but this does mean that should be adopted. Overall superior in terms producing reasonable predictions.