Scholarlink

As the world population grows, varieties of diseases also evolve rapidly. It becomes a challenging task for the healthcare systems to identify and predict the diseases in the earlier stage to avoid more complications Among all, Alzheimer’s disease (AD) is deadlier since it is one of the main causes of dementia which disturbs the brain in a progressive manner. This causes the elderly people to be inactive by weakening their memory and physical functionality. Early diagnosis and treatment help to recover from the AD successfully and to overcome more complications. Numerous machine learning and deep learning algorithms were in the state-of-art. In this paper, an integrated framework by combining BiLSTM and XGBoost machine learning algorithm with Bayesian optimization technique for predicting the AD is proposed. The performance of the proposed BiLSTM_XGBoost algorithm outperforms the other methods in terms of accuracy rate of 95.62%, Precision of 0.93, recall of 0.93, loss of 0.32, and ROC-AUC rate of 0.79 respectively. The proposed model is trained using the ADNI dataset and identified that the proposed integrated framework has a significant potential to lower the AD drastically.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Leveraging Optimized BiLSTM and XGBoost for Early Detection of Alzheimer’s Disease

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • S. V. Divya,
  • P. Venkadesh,
  • M. Rajalakshmi,
  • C. Berin Jones,
  • S. Yazhini,
  • M. Shify Antolin

摘要

As the world population grows, varieties of diseases also evolve rapidly. It becomes a challenging task for the healthcare systems to identify and predict the diseases in the earlier stage to avoid more complications Among all, Alzheimer’s disease (AD) is deadlier since it is one of the main causes of dementia which disturbs the brain in a progressive manner. This causes the elderly people to be inactive by weakening their memory and physical functionality. Early diagnosis and treatment help to recover from the AD successfully and to overcome more complications. Numerous machine learning and deep learning algorithms were in the state-of-art. In this paper, an integrated framework by combining BiLSTM and XGBoost machine learning algorithm with Bayesian optimization technique for predicting the AD is proposed. The performance of the proposed BiLSTM_XGBoost algorithm outperforms the other methods in terms of accuracy rate of 95.62%, Precision of 0.93, recall of 0.93, loss of 0.32, and ROC-AUC rate of 0.79 respectively. The proposed model is trained using the ADNI dataset and identified that the proposed integrated framework has a significant potential to lower the AD drastically.