http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Rohini M.,Surendran D.,Manoj S. Oswalt 대한전기학회 2022 Journal of Electrical Engineering & Technology Vol.17 No.2
Alzheimer's disease (AD), cerebrovascular disease, Lewy-body disease, and Frontal–temporal degeneration disease are the age-related cognitive impairments that cause dementia. However, AD is the primary cause of dementia that causes brain cell degeneration in the geriatric community. Brain cell degeneration is the crucial cause of AD, due to the abnormal accumulation of indissoluble clumps known as plaques and tangles in the human brain's neurons. Amyloid precursor protein levels and Apolipoprotein -E gene are the biomarkers of AD since it causes accumulations and hence blocks the neuron transport system throughout the body. The early onset of AD includes mild-cognitive impairment (MCI) that progresses to complete dementia. Many related works include AD prediction using clinical modality images and cognitive assessments scores of the individuals but have not addressed comparative genome study for signifi cant subjects. However, there is a lack of aff ordable biomarkers for the eff ective early detection of high-risk individuals. In this study, we utilize one or more features of Magnetic Resonance Imaging (MRI) tests and Apolipoprotein-E genotype sequence that provides more signifi cant biomarkers for the early prediction. The ML classifi ers including Support vector classifi er, Gaussian process, AdaBoost, Random Forest, Decision trees learns the subset of patterns that predicts the AD with gene descriptors from microRNA expression profi le and the profi led gene pattern. These signifi cant multiple gene descriptors provide a supportive prediction methodology that apply genotype strength with the ensemble classifi ers. The fi nal optimal model is given by validation evaluations. The support vector classifi er and Random Forest classifi ers had given consistent results for disease conversion and progression from MRI attributes and had given promising results with the validation that showed accuracy greater than 80% and F1 weighted score of 0.8 in disease classifi cation and prognosis. The experimental results had proven 95% accuracy in the saliency values of APOE isoforms implemented in DragonNN framework that will vary AD pathogenic. Hence particular focus and clinical interventions can be given on Aβ genome dependent subjects that predicts the disease