Stage Wise Classification and prediction of Alzheimer disease using Hybrid Appraoch

Alzheimer’s, the most prevalent cause of dementia, leads to a continuous decline in cognitive functions, social skills, and behavior, affecting individuals’ independent functioning. As per survey 2010 in India around 3.7 million affected by dementia, and this data is expected to twofold by 2030. The associated cost of care is projected to triple, necessitating ongoing support. This issue is a significant societal concern that warrants careful consideration. Prediction and classification of Alzheimer’s disease (AD) are crucial for identifying and addressing it at earlier stages. The disease is typically classified into three stages: early-stage dementia, middle-stage, and late-stage. However, several challenges exist, including limited awareness, diagnostic difficulties, insufficient technologies and treatments, unpredictable progression across stages, and a lack of comprehensive stage-by-stage predictions. This work proposes a hybrid model for predicting, classifying, and stage-wise classifying Alzheimer’s disease. The hybrid model aims to address the challenges associated with the disease, utilizing different features for prediction and classification. The hybrid model used a Random forest, convolutional neural network (CNN), kernel density algorithm (KDA) with hyperparameter tuning for accurate prediction and classification. The proposed hybrid method used Random forest used for basic classification, CNN and kernel density for stage wise classification. The proposed hybrid method achieved an impressive 95.6% accuracy in stage wise classification and prediction of AD.