277 0

Analysis of the association between COVID-19 infection and diabetes and hypertension through ensemble models and classification algorithms

Title
Analysis of the association between COVID-19 infection and diabetes and hypertension through ensemble models and classification algorithms
Author
이유석
Advisor(s)
조인휘
Issue Date
2022. 8
Publisher
한양대학교
Degree
Master
Abstract
COVID-19 is spreading like a trend around the world, paralysis of daily life, and severe human death[1]. In addition, asymptomatic patients are also infected with COVID-19, and many strains of COVID-19 are also being found. Currently, numerous COVID-19 research results and cases are being published. Nowadays, Several medical fields, such as COVID-19, are contributing a lot to human life by combining machine Learning, Deep Learning. More recently, in combination with big data in the medical field, data analysis and patterns for specific diseases have been discovered, helping doctors make decisions. In this paper, We will check whether respondents with common symptoms of COVID-19 have Diabetes & high blood pressure and various diseases, and predict infection rates with various information from respondents such as race and smoking status. Studies have shown that diabetes is a chronic disease of all diseases [2], as diabetes is associated with COVID-19 infection rates [3]. This is because high blood pressure also has many diseases in adults and has many associations with diabetes[4][5]. Therefore, in this paper, we will check the presence or absence of a specific disease among COVID-19 infected people and analyze whether it is related to COVID-19 infection or disease. Through this, the main purpose is to find the relationship between a specific disease and the COVID-19 infection rate and to inform them that the risk of infection is high when they have a specific disease. Through the machine learning package provided by Pyspark, we will create a pipeline to properly process the data, combine the two models with the highest f-score to create an ensemble model, and compare them with the classification algorithm to find the best model. As a result, 92% of respondents with diabetes hypertension can be confirmed to be infected with COVID-19, and the infection rate is 12% higher than that of other common disease respondents. In conclusion, if you have diabetes and hypertension, the infection rate is higher than that of the general person even if you do not have general COVID-19 symptoms.
URI
http://hanyang.dcollection.net/common/orgView/200000627559https://repository.hanyang.ac.kr/handle/20.500.11754/174212
Appears in Collections:
GRADUATE SCHOOL[S](대학원) > COMPUTER SCIENCE(컴퓨터·소프트웨어학과) > Theses (Master)
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE