An Optimal Feature Selection Based Ensemble Classification Model to Predict AD Disease Pattern

Abstract

In recent years machine learning acting an essential role in disease pattern classification role and the collection of numerical methods is proven the effectiveness of Alzheimer's disease prediction. Alzheimer's disease (AD) might be a neurodegenerative disorder of hesitant cause and pathogenesis that mainly affects older adults and is that the commonest details for dementia. The original clinical demonstration of AD is discriminating memory impairment and while treatments are accessible to advance some symptoms, there's no cure at present available. Machine learning algorithms are suitable and have gained major attention to classify the AD pattern. We offer a comprehensive synopsis and analysis of the foremost recent research on this topic. In recent years, many supervised classification models are applied and gain significant accuracy on classification, including both linear and Nonlinear SVM, Decision Tree, Random forest, Ada Boost models. The major intention of this article is early detection of the MCI and AD using machine learning classification models, including both supervised and unsupervised models. This paper proposed an optimistic feature selection-based Hybrid Ensemble classification model on ADNI and OASIS data sets to the early recognition of Alzheimer's disease severity. Experiment results are shown better efficiency than existing classifications models. The efficiency of the proposed model is in terms of accuracy, precession, recall, and f1-score.