id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_wfsjha4f5nb7vih3n7vddj25zu	Vitaly Schetinin	Prediction of survival probabilities with Bayesian Decision Trees	2013	33	.pdf	application/pdf	10522	857	63	Practitioners use Trauma and Injury Severity Score (TRISS) models for predicting the survival probability of an injured patient. propose the Bayesian method for estimating the predictive density and show We tested the proposed Bayesian method on a set of patients registered In [33], conventional and Bayesian logistic random effects regression models were compared for predicting outcomes on a data set of 8,509 patients When Bayesian inference is employed for predictions, it is typically assumed that there exist a number of models which can appropriately approximate the relationship between predictor variables and output variable (or Figure 1: Observed and predicted survival probabilities for patients with different numbers of injuries. of the predictive density in terms of entropy of model mixing was significantly higher than that for the proposed method. Based on a regression model, the TRISS method cannot provide the estimates of predictive probability density that is required to evaluate confidence	./cache/work_wfsjha4f5nb7vih3n7vddj25zu.pdf	./txt/work_wfsjha4f5nb7vih3n7vddj25zu.txt