ABSTRACT
This chapter discusses linear regression and some other techniques for determining initial estimates. Modeling can be separated into two classifications, linear regression and nonlinear regression. These two types of regression can be classified according to the methods used to determine the best parameter values. In the case of linear regression, explicit equations for the slope and the intercept can be developed. By assuming a one-compartment pharmacokinetic model for disposition, J. G. Wagner and E. Nelson developed a specific deconvolution method. The Wagner-Nelson and the Loo-Riegelman methods are derived using a compartmental, one- or two-compartment model. Ideally, data points should be collected until drug concentrations become insignificant. More general approaches can be developed using deconvolution techniques. Basically, a deconvolution technique allows the analyst to mathematically subtract information obtained after intravenous administration from data collected after oral, or other, extravascular administration.
