Imputasi Data Menggunakan Teknik Statistik Hot-Deck Untuk Klasifikasi Penyakit Menular HIV Dengan Metode Random Forest

Abstract

Infectious diseases caused by microorganisms, such as the Human Immunodeficiency Virus (HIV), remain a significant global health issue. In health data management, the problem of missing values frequently arises, potentially compromising data quality and affecting analysis results and classification processes. Therefore, appropriate methods for handling missing values are essential to maintain data quality. This study aims to implement the Hot-deck Imputation method to address missing values in an HIV dataset and to evaluate classification performance using the Random Forest algorithm. The dataset was obtained from Harvard Dataverse and split into two subsets: 80% for training and 20% for testing. Model performance was evaluated using several metrics: Accuracy, Precision, Recall, and F1-Score. The evaluation results demonstrate that the Hot-deck Imputation method successfully handled missing values, enabling the dataset to be used for classification with the Random Forest algorithm. Testing results showed an Accuracy of 56.07%, Precision of 54.86%, Recall of 63.58%, and an F1-Score of 58.90%. These results serve as a pre-processing step for handling missing values prior to the classification process for the HIV dataset.