Improving the accuracy in prediction of patient admission in emergency ward using naive bayes compared with logistic regression
G. Monish Kumar, P. Jesu Jayarin · 2024
The objective of this research is to enhance precision in forecasting patient admission utilizing the novel Naive Bayes (NB) model versus logistic regression (LR). Patient admission prognosis in the emergency department is executed employing the novel NB and LR models with 10 iterations, calculated through ClinCalc with Gpower set at 0.1 and alpha at 0.05. The dataset comprises 3,18,438 patient entries, where 15,757 rows and 56 columns serve as features. The evaluation of patient admission accuracy and efficacy is derived from the outcomes. The mean accuracy for patient admission management using the novel NB model stands notably high at 93%, surpassing LR 91%. The statistical significance of the accuracy is 0.060 (p > 0.05) according to the independent sample t-test, suggesting no substantial difference between the two algorithms. The mean accuracy of the novel NB model outperforms LR in hospital patient admission prediction.