Tensor Based Multi-linear Feature Selection Models to Predict Alzheimer Disease

Main Article Content

Sai Sindhuri Nasina, Prof. A. Rama Mohan Reddy

Abstract

Alzheimer's disease is one of the progressing diseases that affect the brain cells cruelly. It triggers nerve cell death and brain tissue loss. It also starts slowly and worsens overtime. The signs of Alzheimer's disease depend on the severity of the condition from person to person. It displays behavioral effects such as poor speech, lack of memory, longer time to complete daily tasks and change of mood and behavior. It cannot be fixed if the condition worsens over time. This means that at the earliest point it should be detected and the patient should be cared with a normal life alone. Deep learning algorithms have wonderful success in detecting complex patterns in vast quantities of high-dimensional medical imaging knowledge over traditional machine learning algorithms. Therefore, a great deal of attention was paid lately to applying profound learning to medical diagnosis. The goal of this investigation is to identify the various phases of Alzheimer's disease from the Magnetic Resonance Imaging (MRI) images by using the Multi-Linear Singular Value Decomposition (MLSVD) and the tensors inspired Multi-Linear Principal Component Analysis (MLPCA) models. An ADNI dataset is experimented and the findings demonstrate that excellent precision has been obtained by the proposed models.


 

Article Details

Section
Articles