Borderline-DEMNET for Multi-Class Alzheimer's Disease Classification

Alzheimer's Disease (AD) is a neurological condition characterized by the gradual shrinkage and death of brain cells. It is a common form of Dementia that leads to cognitive impairments such as problems with thinking, reasoning, and remembering. Detecting and predicting AD accurately is crucial due to its significant impact on mortality. This study presents a novel framework called Borderline-DEMNET, which addresses the Class Imbalance issue using the Borderline-SMOTE technique. Borderline-DEMNET is an extension of the DEMNET model described in related works. The experimental results demonstrate that the proposed framework achieves a testing accuracy of 99.14%, surpassing the accuracy of the base model. Furthermore, it outperforms other multi-class classification methods for AD in terms of accuracy, F1-score, recall, and precision. These findings indicate that the suggested framework exhibits superior performance across multiple evaluation metrics.