Optimasi Decision Tree menggunakan Pendekatan Ensemble Bagging untuk Klasifikasi Klinis Hipertensi

Abstrak

Hypertension is a common health problem with a high risk of serious complications, so it is important to detect it early. This study is an experimental quantitative study that aims to improve the accuracy of the hypertension classification model by optimizing the Decision tree algorithm using the Bagging ensemble method. The aspects studied include the effect of data balancing techniques on the performance of classification models in detecting the risk of hypertension. The dataset used consisted of 2,655 medical records of primary health care patients. The research stages included data pre-processing with imputation, normalization, and balancing with SMOTE, Decision Tree and Bagging modeling, and model evaluation using 5-fold cross-validation, Confusion Matrix, ROC-AUC, and paired t-test. The results showed that the application of Bagging significantly improved the performance of the Decision Tree (p < 0.05) with an accuracy of 96.88%, precision of 95.56%, recall of 98.32%, F1-score of 96.93%, and AUC of 0.996. This improvement indicates that the Bagging method is able to reduce model variance and produce more stable predictions on imbalanced clinical data. These findings contribute to the development of a classification model based on real clinical data, which also opens up opportunities for the development of machine learning-based Clinical Decision Support System (CDSS) for early detection of hypertension in primary health facilities.

Identifier

• DOI: https://doi.org/10.29408/edumatic.v9i3.32400
• OpenAlex: https://openalex.org/W7124748796