Predicting the Impact of Socio-Demographic Risk Factors on COVID-19 Based on Hybrid ANN-CNN Model

Global health has been greatly influenced by the COVID-19 pandemic, especially in low- and middle-income nations like Nigeria. Despite the catastrophic effects of the pandemic, little is known about how sociodemographic risk factors that affects the number of COVID-19 infections and deaths in Nigeri...

Full description

Saved in:

Bibliographic Details
Published in:	2023 First International Conference on the Advancements of Artificial Intelligence in African Context (AAIAC) pp. 1 - 8
Main Authors:	Sani, Yahaya Mohammed, Davou Pam, Benjamin, Adamu, Mamman
Format:	Conference Proceeding
Language:	English
Published:	IEEE 15-11-2023
Subjects:	Artificial Neural Network Convolutional Neural Network COVID-19 Hybrid ANN-CNN Least Absolute Shrinkage and Selection Operator Long Short Term Memory and COVID-19 Machine learning Mathematical models Multiple Linear Regression Pandemics Predictive models Rivers Socio-demographic Training
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Global health has been greatly influenced by the COVID-19 pandemic, especially in low- and middle-income nations like Nigeria. Despite the catastrophic effects of the pandemic, little is known about how sociodemographic risk factors that affects the number of COVID-19 infections and deaths in Nigeria. Using Spearman heat map correlation analysis, this study examined these parameters and developed a hybrid ANN-CNN model to forecast the influence of sociodemographic characteristics against COVID-19 confirmed cases and mortality cases in Nigeria. The Nigerian COVID-19 confirmed and death cases data from May 1, 2020, to April 30, 2021, as well as sociodemographic risk factor statistics, were the datasets used in this study. The experiment was completed by training and testing the models, and based on MAEs and RMSEs models performance evaluation metrics, the developed Hybrid ANN-CNN model outperformed the other five state-of-the-art machine learning models involving Multiple Linear Regression (MLR), Artificial Neural Network (ANN), Convolutional Neural Network (CNN), Long Short Term Memory (LSTM), and Least Absolute Shrinkage and Selection Operator (LASSO). With mean absolute errors of (0.0157) and (0.0181) for confirmed and death cases, respectively, the developed Hybrid ANN-CNN model outperformed the others. Similarly, with RMSEs of (0.0842) and (0.0813) for confirmed and death cases, respectively, the developed Hybrid ANN-CNN model fared better than other models. The developed Hybrid ANN-CNN model can be helpful in tracking and containing pandemic outbreaks, both in the present and the future.
DOI:	10.1109/AAIAC60008.2023.10465311