|
Abstract:
|
In this talk, we propose several dose-time-effect models, which combine the median effect model with the Gompertzian growth curve model. We apply the least squares method and the transformation-both-sides to estimate the parameters in the models. The Akaike information criterion (AIC) is used to select a parsimonious model. We fit the dose-time-effect models to an in-vitro experiment data set to illustrate the proposed procedures for estimating parameters and selecting models. For the case study, the selected model fits the experimental data quite well. By comparing different dose-time-effect models with the selected model, we found that the cells killed by treatment still contributed to the total cell count. By accounting for these killed cells in our models, the relationship between dose and the fraction of cells surviving treatment can be more accurately elucidated.
|
