<![CDATA[This study compares the performance of the Decision Tree and Extreme Gradient Boosting (XGBoost) algorithms in classifying HIV/AIDS infection status. A quantitative experimental design was employed using a secondary dataset of 2,139 records and 23 attributes obtained from an open-source platform. Data preprocessing included checking for missing values, removing duplicates, detecting and handling outliers with the Interquartile Range (IQR) method, and applying feature scaling. The models were trained and tested with three data split ratios (70:30, 80:20, and 90:10). Evaluation metrics comprised accuracy, precision, recall, and F1-score derived from the confusion matrix.The results show that XGBoost achieved the highest performance, reaching 99.16 % accuracy, 98.17 % precision, 99.16 % recall, and 99.16 % F1-score with a 90:10 data split. In comparison, the Decision Tree achieved a maximum accuracy of 95 % with an F1-score of approximately 95 % under the same conditions. These findings confirm that XGBoost consistently outperforms the Decision Tree in accuracy and generalization across all data-split scenarios. This research concludes that XGBoost is more suitable for developing data-driven decision support systems for HIV/AIDS infection detection.]]>
Copyrights © 2025
