A BAYES COMPARISON OF TWO DIFFERENT CANCER THERAPIES UNDER THE ASSUMPTION OF WEIBULL SURVIVAL MODEL OR ITS SUBFAMILY

  • Puja Arora Department of Statistics DST Centre for Interdisciplinary Mathematical Sciences Banaras Hindu University, Varanasi-221 005, India
  • S.K. Upadhyay Department of Statistics DST Centre for Interdisciplinary Mathematical Sciences Banaras Hindu University, Varanasi-221 005, India
Keywords: Weibull Distribution, Shape Parameter, Exponential Distribution, Survival Function, Noninformative Prior, p-value, Chi-square Discrepancy Measure, Bayes Information Criterion

Abstract

The paper considers a group of patients suffering from leukemia B non-small lung cancer. Such patients are generally suggested to undergo for either radiotherapy or chemotherapy followed by radiotherapy. The objective of the paper is to compare the two therapies based on survival functions of the patients assuming Weibull survival model for each therapy. The paper further examines the feasibility of a subfamily of Weibull model, namely the exponential distribution, for a date set available from a clinical trial experiment. This feasibility is judged based on Bayes information criterion by comparing the Weibull model with its subfamily. The model compatibility study with the data based on posterior p-values has also been given to ensure the suitability of the two models. Finally, the recommendations are made accordingly.

 

Downloads

Download data is not yet available.
Published
2008-06-30
Section
Articles