Related Experiment Video
Updated: Nov 21, 2025
04:17
DNA Virus Detection System Based on RPA-CRISPR/Cas12a-SPM and Deep Learning
Published on: May 10, 2024
1.2K
Forecasting of COVID-19 cases using deep learning models: Is it reliable and practically significant?
Jayanthi Devaraj1, Rajvikram Madurai Elavarasan2, Rishi Pugazhendhi3
1Department of Information Technology, Sri Venkateswara College of Engineering, Chennai 602117, India.
Results in Physics
|January 19, 2021
Summary
Artificial Intelligence (AI) models, including Stacked Long Short-Term Memory (SLSTM), accurately predict COVID-19 cases. The SLSTM model achieved less than 2% error, aiding pandemic management and supporting Sustainable Development Goals.
Area of Science:
- Computational epidemiology and public health
- Artificial Intelligence in healthcare
- Time series forecasting for infectious diseases
Background:
- The COVID-19 pandemic poses a significant threat to global health and economic stability.
- Accurate prediction of COVID-19 cases is crucial for effective pandemic response and resource allocation.
- Traditional methods are insufficient for the dynamic nature of pandemic spread.
Purpose of the Study:
- To evaluate the efficacy of various Artificial Intelligence (AI) and deep learning models for predicting global and country-specific COVID-19 cases.
- To identify the most accurate and reliable forecasting model for short-term, medium-term, and long-term predictions.
- To analyze the correlation between environmental factors and COVID-19 transmission.
Main Methods:
- Data pre-processing and feature extraction from real-world COVID-19 datasets.
- Implementation and comparison of Auto-Regressive Integrated Moving Average (ARIMA), Long Short-Term Memory (LSTM), Stacked LSTM (SLSTM), and Prophet models for case prediction.
- Multivariate LSTM models were used for long-term forecasting, alongside statistical hypothesis and correlation analyses incorporating environmental and demographic data.
Main Results:
- The Stacked LSTM (SLSTM) algorithm demonstrated superior accuracy, achieving prediction errors below 2% across key performance metrics.
- SLSTM outperformed ARIMA, standard LSTM, and Prophet models in forecasting cumulative confirmed, death, and recovered COVID-19 cases.
- Detailed country-specific (India) and city-specific (Chennai) analyses were performed, alongside correlation analysis with environmental factors.
Conclusions:
- Stacked LSTM is a highly accurate and reliable model for predicting COVID-19 cases, offering significant advantages over other time series methods.
- AI-driven predictions are vital for pandemic management, enabling better scenario planning and resource optimization.
- Accurate forecasting supports the achievement of Sustainable Development Goals by mitigating the pandemic's societal and economic impacts.
