Diagnosis of Alzheimer Disease using Machine Learning Approaches

Abstract

This article applies the machine learning paramount techniques for the early detection and effective diagnosis of severe Alzheimer’s disease (AD). AD is neurodegenerative chronic disease and often essential to detect at the early stage. It is vital to diagnose the disease in initial stages for more effective and beneficial treatment. Machine learning is becoming the booming area and shows the remarkable achievement in the present, advance and crucial decision making. Medical diagnosis is one of the crucial areas that have paramount importance where various learning algorithms can be contributed for the improvement in disease diagnosis. Due to the evolution of computation technology, the generation of data is increased exponentially, especially in medical field. To cope up this problem, this paper elaborates and tests various approaches of machine learning for AD diagnosis. Using oasis_longitudinal MRI data, this study train various models of machine learning to detect patients suffering from the AD. Along with the machine learning, this letter addressed the deep learning comprehensive overview. With various learning approaches such like binary classifier named logistic regression (LR), support vector machine (SVM), hierarchical decision tree (DT), ensemble random forest (RF), and boosting adaboost, experimental result is analyzed in terms of accuracy, recall, and AUC (Area Under Curve). The results obtained concluded that random forest and adaboost achieve higher accuracy, together with random forest also able to get higher recall and AUC. The minimum time to accomplish the classification task is taken by decision tree that is 68.788ms. This document result would be helpful to strengthen the idea and concept of applying the learning algorithms in disease detection at early stage.