Prediction of Sepsis Using Light Gradient-Boosting Machine Classifier in Comparison with Adaboost Classifier Based on Accuracy
Main Article Content
Article Sidebar
Abstract
This study introduces a method to forecast sepsis employing the innovative LightGBM classifier model,
juxtaposing its improved accuracy against the Adaboost Classifier model. The dataset was sourced from
PhysioNet/Computing in Cardiology Challenge 2019's training set. The G power software informed the
sample size decision, suggesting 10 participants for each group, adopting a pretest power of 80%. A 95%
confidence interval was applied, and a significance level was established at 0.05%. Remarkably, the
LightGBM Technique achieved 96.41% accuracy, surpassing the AdaBoost Classifier's 77.58%. A significant
difference was observed between the two, evidenced by a P value of 0.019. In conclusion, the Light GradientBoosting Machine classifier offers superior accuracy in predicting sepsis events
How to Cite
Article Details
Machine Learning, Adaboost Classifier, Innovative Novel LightGBM Technique