<![CDATA[Dementia is a progressive neurodegenerative disorder that leads to cognitive decline and significantly affects patients' quality of life. Early detection is crucial for determining appropriate medical interventions and slowing disease progression. This study aims to develop a machine learning-based dementia prediction model and compare the performance of three algorithms: Support Vector Machine (SVM), Random Forest (RF), and XGBoost. The dataset, obtained from the Kaggle platform, consists of 373 MRI-based patient records categorized into three diagnosis groups: Converted, Demented, and Nondemented. To address class imbalance, the Synthetic Minority Oversampling Technique (SMOTE) was applied. Experimental results show that the XGBoost algorithm achieved the best performance, with an accuracy of 93.86%, precision of 94%, recall of 94%, and F1-score of 94%, outperforming SVM and Random Forest. The application of SMOTE improved the model’s sensitivity to minority classes. The combination of XGBoost and SMOTE demonstrates high accuracy in dementia prediction and holds potential for integration into clinical decision support systems (CDSS) to assist early diagnosis.]]>
Copyrights © 2025
