A comparative study on various machine learning approaches for the detection of Alzheimer disease (original) (raw)

Alzheimer's disease is the most common cause of Dementia. It accelerates the degeneration of brain cells and the progression of memory loss. Identifying and predicting Alzheimer's disease in its early stages is hard. A machine learning system that can predict the disease can solve this problem. High-dimensional data analysis is a massive problem for engineers and researchers in Machine learning. A simple and effective way to solve this problem is to use feature selection to eliminate redundant and useless data. This study aims to determine how accurate different machine learning methods are at diagnosing Alzheimer's disease with help of the feature selection method. Therefore, we utilize the Open Access Series of Imaging Studies dataset in several Machine learning models to accurately detect and predict Alzheimer's disease. We use the wrapper feature selection method in the proposed work to choose the minimal attribute set from the textual records. These approaches include Random Forest and Support vector machine as well as Decision Tree, XGBoost and Voting. We also utilize AdaBoost and Gradient Boosting on the selected data to classify Alzheimer's disease from the Alzheimer's disease vs. Normal disease dataset. As part of the analysis of the results, we found that the machine learning with feature selection methods provides better result as compared to machine learning without feature selection methods.