Online Program

Return to main conference page

All Times EDT

Thursday, October 7
Thu, Oct 7, 2:45 PM - 4:00 PM
Speed Session

Using Artificial Neural Network to Model Factors That Influence the Action of a Person Going to Coronavirus Testing (309916)

*Sheila Rutto, The University of Texas Rio Grande Valley 

Keywords: COVID-19, coronavirus, ANN, AUC

COVID-19 became the first pandemic that brought the entire world to an abrupt and unexpected halt. Nonetheless, due to the urgency of the matter, rapid studies were conducted to understand its cause, risk factors, symptoms, and mitigation that, can be taken against coronavirus. The US Department of Health and Human Services is leading in advocating for Americans to be tested for the coronavirus as it is an essential step in fighting the virus. It assists in identifying positive cases, isolated to prevent the spread of the virus. If identification is known early enough, it can help reduce fatal cases if not properly treated on time. Fighting against coronavirus testing is vital but, it is an individual choice to take a test or not. This study attempts to establish a prediction model using the artificial neural network (ANN) technique to establish factors that can influence someone to take the COVID-19 test. Using COVID-19 survey data from The Understanding American Study website, several models fitted using different machine learning techniques. ANN model emerged the best in terms of model performance, with an AUC value of 0.9525. The model was trained and tested using 211 and 88 observations, respectively. Given such high performance, the model may be, used in predicting if an individual can take a corona test or not, which can be helpful information for policymakers.