Automated Classification of Alzheimer’s Disease using Deep Neural Network (DNN) by Random Forest Feature Elimination

Determining Alzheimer's Disease (AD) in its early stages is very important to prepare proper care for the patient. In this study, we aimed to create fast and accurate automated classification system to determine AD with the minimum data collected from the patient. Magnetic Resonance Imaging (MRI) is widely used to diagnose AD. When the cost of the technique and risks of the procedures are considered, there is a need for different solutions. With the availability of neural network chips, it is even possible to build portable devices for Alzheimer's detection. We propose fast and successful method to detect Alzheimer using Deep Neural Network (DNN). To reduce the complexity of the algorithm, Random Forest method was used to eliminate some of the features. Success of Random Forest to eliminate the features and success of DNN to detect AD are discussed.