Ensemble Methods for Early Prediction of Dementia

Dementia is a progressive cognitive disorder that primarily affects older adults. The patients affected by Dementia experience memory loss, reasoning impairment, and behavioural changes. This condition significantly impacts the daily life and independent activities of the individuals. Although currently available medications cannot reverse or stop the neurodegeneration processes they can temporarily improve or maintain its cognitive functions in the early to moderate phases of the disease. However, early detection of the disease may delay its progression. The paper proposes a systematic approach by applying ensemble methods to predict dementia risk. The algorithms used included XGBoost (XGB), Gradient Boost (GB), Random Forest (RF), Voting Classifier (VC) includes Naïve Bayes (NB) and Support Vector Machine (SVM). The classification algorithms have been compared based on evaluation metrics like accuracy, precision, recall, and f1 score. The model developed achieved an accuracy of 97% and outperformed other methods on the same dataset. The final model enables the early detection of dementia, which is vital for slowing down the neurological decline associated with the disease. This approach is beneficial for underserved populations with limited access to healthcare providers.