TT self-weighted Deep-AD3-Net: An AD stage and risk prediction

Abstract

ABSTRACT An age-related disease with memory deficits and cognitive decline is termed Alzheimer's Disease (AD). For AD prevention, the characterization of at-risk states and timely detection of AD is imperative. So, in this work, an efficient AD stage and risk prediction model is proposed. Afterward, by utilizing the Recursive Hypothesis-Creation Algorithm (RHCA), the crucial variables are selected and ranked via the tailored metric ranking techniques. Then, by employing the relevant variables obtained from the original ADAS Cog-13, the risk prediction score is calculated utilizing the True True Self-Weighting Mechanism (TT-SWM) and the ventricle volume weight assignment process. Next, by utilizing the Queue-Boltsman-Constant-Sphere (QBCS) approach, the volume and hippocampus area of the brain of an AD-affected person and the rate of shrinkage in the brain are gauged. At last, with the bilateral hippocampus, resting-state functional Magnetic Resonance Imaging (rs-fMRI) is evaluated for the whole brain utilizing the Phylogenetic Method (PM). In this, via the creation of one Gray-Level Co-occurrence Matrix (GLCM), the bilateral hippocampus features are obtained. Next, utilizing a Genetic Algorithm (GA), the features are extracted and inputted into the Deep-AD3-Net classifier for the classification of diverse AD stages.