Kevin Zhao, Kevin Qi, Daniel Che, M. Shalaginov, Tingying Helen Zeng, Megan Pugach-Gordon
{"title":"An Advanced Machine Learning Algorithm For Intensive Care Decisions Of COVID-19 Patients","authors":"Kevin Zhao, Kevin Qi, Daniel Che, M. Shalaginov, Tingying Helen Zeng, Megan Pugach-Gordon","doi":"10.1109/BioSMART54244.2021.9677790","DOIUrl":null,"url":null,"abstract":"COVID-19, an infectious respiratory disease, is a global health crisis and severely taxed healthcare systems. The SARS-CoV-2 virus damages lungs and other vital organs and even causes acute respiratory distress syndrome (ARDS). Currently, intensive care, including supplemental oxygen and ventilation, is used to treat severe cases. In this project, a Machine Learning algorithm was developed to predict intensive care needs for patients in the early stage of Covid-19. An advanced convolutional neural network (CNN) model was trained for image classification based on patient chest x-rays. After studying and comparing the performance of several advanced models, including Inception V3,ResNet50, Xception, EfficientNetB0, EfficientNetB7 and VGG16, It is identified that Inception V3showed the highest accuracy of the prediction. Based on Inception V3,an algorithm that demonstrates the highest accuracy of over 99% on both validation and testing datasets has been developed. The algorithm accurately makes predictions for which patients need immediate intensive care, so as to help the COVID19 patients” recovery and save more lives.","PeriodicalId":286026,"journal":{"name":"2021 4th International Conference on Bio-Engineering for Smart Technologies (BioSMART)","volume":"106 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 4th International Conference on Bio-Engineering for Smart Technologies (BioSMART)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/BioSMART54244.2021.9677790","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
COVID-19, an infectious respiratory disease, is a global health crisis and severely taxed healthcare systems. The SARS-CoV-2 virus damages lungs and other vital organs and even causes acute respiratory distress syndrome (ARDS). Currently, intensive care, including supplemental oxygen and ventilation, is used to treat severe cases. In this project, a Machine Learning algorithm was developed to predict intensive care needs for patients in the early stage of Covid-19. An advanced convolutional neural network (CNN) model was trained for image classification based on patient chest x-rays. After studying and comparing the performance of several advanced models, including Inception V3,ResNet50, Xception, EfficientNetB0, EfficientNetB7 and VGG16, It is identified that Inception V3showed the highest accuracy of the prediction. Based on Inception V3,an algorithm that demonstrates the highest accuracy of over 99% on both validation and testing datasets has been developed. The algorithm accurately makes predictions for which patients need immediate intensive care, so as to help the COVID19 patients” recovery and save more lives.