Survey on regression analysis of photoplethysmography using machine learning
Autor: | G. Valarmathi, K. Sivasankari, Su. Suganthi, S. Kavitha, R. Prabha, V. Subashini, R. Janaki |
---|---|
Rok vydání: | 2021 |
Předmět: |
010302 applied physics
medicine.medical_specialty business.industry Regression analysis 02 engineering and technology 021001 nanoscience & nanotechnology medicine.disease 01 natural sciences Prehypertension Normal tension Photoplethysmogram Internal medicine 0103 physical sciences medicine Cardiology sense organs 0210 nano-technology business Stroke |
Zdroj: | Materials Today: Proceedings. 46:3743-3748 |
ISSN: | 2214-7853 |
Popis: | Hypertension is most common chronic cardiovascular disease (CVD). Diagnosing the Hyper tension initially place a vital role in preventing the occurrence of CVDs and stroke. Early screening can be done using PPG. photoplethysmography (PPG). Various morphological highlights of PPG were studied and assessment was done to eliminate variations. Subsidiary waves were characterized and removed. Six kinds of highlight choice techniques were picked to screen and assess these PPG morphological highlights. The ideal highlights can be extensively broke down corresponding to the physiological procedures of the cardiovascular circulatory framework. Especially, the natural connection and physiological importance between the development procedure of systolic circulatory strain (SBP) and PPG morphology highlights were examined inside and out. An assortment of direct and nonlinear arrangement models were set up for the correlation preliminaries. The F1 scores for the normotension versus pre hypertension, normal tension and prehypertension versus hypertension, and normotension versus hypertension were 72.97%, 81.82%, and 92.31%, respectively. In outline, this examination set up a PPG trademark investigation model and set up the natural connection among SBP and PPG qualities. At last, the risk of hypertension at various phases were analysed and compared based on optimal features. |
Databáze: | OpenAIRE |
Externí odkaz: |