Incorporating Prior Beliefs into the Construction of Prior Distribution of the Parameters: A: A Survival Analysis Approach

Bayesian inference requires the combination of prior experience and the observed data (in the form of the likelihood). There is the need to discuss the specification of prior information about the model parameters. This research has built a structure for the joint prior distribution depending on the type of covariate in the context of a general linear model. Some issues associated with prior elicitation and some ways of incorporating our prior beliefs into the construction of the prior distribution of the parameters of the survival model are discussed in this research. This study also presents a way of specifying the covariance between parameters by thinking in terms of the coefficient of determination to find the correlation between the parameters. The research has given a direct and meaningful approach which reduces the burden on experts.