The Application of Supervised and Unsupervised Computational Predictive Models to Simulate the COVID19 Pandemic

Resource type
Title
The Application of Supervised and Unsupervised Computational Predictive Models to Simulate the COVID19 Pandemic
Abstract
The application of different tools for predicting COVID19 cases spreading has been widely considered during the pandemic. Comparing different approaches is essential to analyze performance and the practical support they can provide for the current pandemic management. This work proposes using the susceptible-exposed-asymptomatic but infectious-symptomatic and infectious-recovered-deceased (SEAIRD) model for different learning models. The first analysis considers an unsupervised prediction, based directly on the epidemiologic compartmental model. After that, two supervised learning models are considered integrating computational intelligence techniques and control engineering: the fuzzy-PID and the wavelet-ANN-PID models. The purpose is to compare different predictor strategies to validate a viable predictive control system for the COVID19 relevant epidemiologic time series. For each model, after setting the initial conditions for each parameter, the prediction performance is calculated based on the presented data. The use of PID controllers is justified to avoid divergence in the system when the learning process is conducted. The wavelet neural network solution is considered here because of its rapid convergence rate. The proposed solutions are dynamic and can be adjusted and corrected in real time, according to the output error. The results are presented in each subsection of the chapter.
Book Title
Epidemic Analytics for Decision Supports in COVID19 Crisis
Place
Cham
Publisher
Springer International Publishing
Date
2022
Pages
103-139
Language
en
ISBN
978-3-030-95281-5
Accessed
9/21/22, 2:33 AM
Library Catalog
Springer Link
Citation
Fong, S. J., Marques, J. A. L., Li, G., Dey, N., Crespo, R. G., Herrera-Viedma, E., Gois, F. N. B., & Neto, J. X. (2022). The Application of Supervised and Unsupervised Computational Predictive Models to Simulate the COVID19 Pandemic. In J. A. L. Marques & S. J. Fong (Eds.), Epidemic Analytics for Decision Supports in COVID19 Crisis (pp. 103–139). Springer International Publishing. https://doi.org/10.1007/978-3-030-95281-5_5