Comparison between logistic regression and machine learning algorithms on survival prediction of traumatic brain injuries.

摘要: Abstract Purpose To compare twenty-two machine learning (ML) models against logistic regression on survival prediction in severe traumatic brain injury (STBI) patients a single center study. Materials and methods Data was collected from STBI admitted to the Sichuan Provincial People's Hospital between December 2009 November 2011. Twenty-two were tested, their predictive performance compared with (LR) model. Receiver operating characteristics (ROC), area under curve (AUC), accuracy, F-score, precision, recall Decision Curve Analysis (DCA) used as metrics. Results A total of 117 enrolled. AUC all ML ranged 86.3% 94%. LR 83%, accuracy 88%. The Cubic SVM, Quadratic SVM Linear higher than that LR. precision ratio 95% 91%, both lower most models. F-Score 0.93, which only slightly better Discriminant Discriminant. Conclusions selected have capabilities comparable classical model for outcome patients. Of these, performed significantly