A proposed network of an effective deep belief for the recognition of Alzheimer

University of Information Technology &Communications, College of Business Informatics Technology, Business Information Technology Department , Baghdad, Iraq

الملخص

Alzheimer(A.Z) is a gradually advancing condition that leads to the degeneration of brain cells. rendering an individual unable of independent functioning. During the initial stages of AZ development, an individual may forget prior talks or the occurrence of an incident. Subsequently, there may be a substantial decline in memory, rendering the individual unable of doing daily tasks . Therefore, this study aims to differentiate between patients with AZ and those in the normal control (NC) group by employing magnetic resonance imaging (MRI). Four phases were used in our study: collected (420)subjects as dataset , preprocessing with a 2D Gaussian filter, and feature extraction by using deep neural networks , while last step is classification . Machine learning algorithms are used to determine if the subject are demented or not .  In this study , applied Random Forest and Naive Bayesian

methods as classifier . For analysis purpose , (WEKA.) tool is utilized .The experimental  results show that accuracy was 83 %with  Random forest while 79% with Naive Bayesian