KLASIFIKASI DIAGNOSA PENYAKIT ANEMIA BERDASARKAN DATA HEMATOLOGI MENGGUNAKAN EXTREME GRADIENT BOOSTING (XGBOOST)

Anemia is a medical condition characterized by low levels of hemoglobin, hematocrit, and red blood cells, which can reduce productivity and quality of life. This study aims to develop an anemia diagnosis classification model based on hematological data using the eXtreme Gradient Boosting (XGBoost) algorithm with hyperparameter tuning and ADASYN oversampling technique. The research dataset consists of 500 medical records from Pelengkap Medical Center Hospital in Jombang, featuring HGB, PCV, RBC, and MCV parameters. Data preprocessing included encoding, normalization, and 70:30 dataset splitting. Results show that the XGBoost model with GridSearchCV hyperparameter tuning achieved 92.66% accuracy. The application of ADASYN 350 samples improved accuracy to 93.33%, while ADASYN 500 unexpectedly decreased model performance (91.33% accuracy). These findings indicate that the combination of XGBoost with ADASYN 350 represents an optimal approach for hematology based anemia classification.