The Data Dependence in Modeling Rat's Tumor Growth Data

• Activity Number: 360
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Chong Yau Fu*+ and Yueh-Hsing Ou and Shih-Hua Liu
• Companies: National Yang Ming University and Institute of Bioonology in Medicine and National Yunlin University of Science and Technology
• Keywords: two-stages ; , multilevel model ; data -rich data

Abstract

The tumor growth is dedicated as an exponential or a Gompertz growth model and interpreted by differential equation, where the sample size is small (e.g. 4~ 12) collected via time and larger than the sample size, called data-rich situation. Based on this special data, the method that each rat fits an individual curve and then aggregates those estimators together is a natural concerned model. This method is denoted as standard two-stages (STS) and modified as Global two-stages (GTS), which are commonly used in the pharmacokinetic model. Except sparse individuals, the data-rich data is correlated data that a model characterizing the data correlation is inquired. Where, mixed model or multilevel model (MLM) is concerned with, in the family of generalized liner model. This study intends to investigate those main issues in STS/GTS and MLM and further to figure out their difference.