Lack of Likelihood Ratios May Cause Misinterpretation of Results in Diagnostic Accuracy Studies

This article aims to test whether a machine learning algorithm can give accurate results in septic shock prediction. Therefore, the authors developed models using logistic regression (LR), extreme gradient boosting (XGB), and artificial neural network algorithms. They derived the training dataset from random sampling of 80% of the entire cohort, and the validation dataset comprised the remaining 20% of the cohort.

Source: The Journal of Emergency Medicine - January 1, 2022 Category: Emergency Medicine Authors: Nurettin Özgür Doğan, Kutlu Barış Teke Tags: Letter to the Editor Source Type: research