Tez M. (Yürütücü)
Diğer Uluslararası Fon Programları, 2016 - 2016
Modeling the pharmacokinetic behavior of a particular drug is a valuable tool
in the drug development process. A well-known and commonly used model is twocompartment model which provides good insight into the underlying behavior of
most drugs [1]. The model can be described analytically in the form of a system
of ordinary differential equations. The solution of equation system is nonlinear
form of the model parameters. Furthermore, compartments are correlated across
the equations. In this case, generalized nonlinear least squares (GNLS) estimator is
more efficient than nonlinear least squares (NLS) estimator [2]. The GNLS approach
minimizes the Minkowski metric with respect to model parameters in which the
covariance structure is not ignored.
In this study, estimation of two-compartment model parameters is considered in
case of correlated equations. It is aimed to estimate the unknown model parameters
based on GNLS minimization. For this purpose, genetic algorithm (GA), a wellknown population based search algorithm [3], is used as optimization tool. In order
to reduce the bias of the estimators, Jackknife delete-one algorithm [4] is used. The
suggested approach is applied on simulated data set. It is seen from the results that
bias of parameter estimates is reduced by using Jackknife method which helps to
get statistical inference about the parameters.