Iterative Bayesian Model Averaging For Patients Survival Analysis

Abstract

          Selection of relevant gene for sample classification is common task in most gene expression studies and also one of most challenging issues in field of microarray data analysis. Besides that, most gene selection method has a problem to produce continuous predictors of survival analysis because they fail to account for model uncertainty. Moreover, the limitation of Bayesian Model Averaging algorithm is not suitable to use in situation where the number of predictive variables is greater than the number of samples. As the result, Iterative Bayesian Model Averaging algorithm was implemented to select subset of informative genes for survival analysis on high dimensional microarray data. The Iterative Bayesian Model Averaging method is a multivariate procedure combines the effectiveness of multiple contending models by calculating the weighted average of their posterior probability distributions. In addition, Iterative Bayesian Model Averaging method for patients’ survival analysis which is easy to use, computationally efficient and also provides high prediction accuracy while selecting a small number of predictive genes. In this study, Iterative Bayesian Model Averaging method is compared performance with others exiting method within experimental approach.

Website : http://ir.fsksm.utm.my/53/