PERBANDINGAN MODEL KLASIFIKASI RANDOM FOREST DENGAN RESAMPLING DAN TANPA RESAMPLING PADA PASIEN PENDERITA GAGAL JANTUNG

Rizwan Arisandi (Departement of Computer Science, Faculty of Informatics Engineering, Bina Nusantara University)

Jurnal Gaussian
Vol 12, No 1 (2023): Jurnal Gaussian

Cardiovascular disease that causes heart failure is one of the diseases with the highest mortality rate in the world. Therefore, there is a need for an accurate model to classify heart failure based on clinical information and the lifestyle of patients with the disease, as an alternative solution in administering appropriate drugs. This study compared the classification model of living and deceased heart failure patients based on clinical information and patient lifestyle using the random forest method when using resampling techniques and not using resampling techniques. The results obtained from this study are that the Random Forest model with a combination of the SMOTE and Edited Nearest Neighbors methods is the best model for classifying someone with heart failure as alive or dead. The Random Forest model with a combination of the SMOTE and Edited Nearest Neighbors methods has a high level of classification accuracy in the evaluation model that focuses on recall, namely rf_model_smoteenn can classify 82.96% of patients with living status and 90% of patients with death status.