Sign in Subscribe
COVID-19

Hybrid Modeling Approaches for Predicting COVID-19 Mortality: A Comparative Study Across USA, France, and India

Strategic Insights

— 1 min read

Researchers have published a new study on COVID-19, proposing and assessing three Gaussian Process-based hybrid models for forecasting mortality cases in the USA, France, and India. The study, led by B. Uppalaiah from the Department of Mathematics at GITAM Deemed to be University, used various evaluation metrics to determine the optimal hybrid model for predicting COVID-19 mortality.

Key Takeaways:

  • The researchers proposed and assessed three Gaussian Process-based hybrid models: GP-RBM, GP-LSTM, and GP-CNN, for forecasting COVID-19 mortality cases in the USA, France, and India.
  • The models combined the adaptability of Gaussian Processes with the deep learning functionalities of RBM, LSTM, and CNN to improve prediction precision and uncertainty assessment.
  • The GP-LSTM and GP-CNN hybrid models surpassed the GP-RBM model in accuracy and efficiency, especially with the datasets from France and India.
  • The study used various evaluation metrics, including Symmetric Mean Absolute Percentage Error (sMAPE), Continuous Ranked Probability Score (CRPS), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC), to determine the optimal hybrid model.
  • The findings indicated notable disparities in prediction accuracy, with each model exhibiting both advantages and drawbacks across different epidemiological contexts.
  • The study highlighted the need for choosing suitable hybrid models customized to the unique dynamics of regional COVID-19 data.

Statistics:

  • The study used datasets from the USA, France, and India to evaluate the performance of the hybrid models.
  • The GP-LSTM and GP-CNN hybrid models achieved higher accuracy and efficiency compared to the GP-RBM model, especially with the datasets from France and India.
  • The study reported that the GP-LSTM and GP-CNN hybrid models outperformed the GP-RBM model in terms of Symmetric Mean Absolute Percentage Error (sMAPE), with values of 0.12 and 0.15, respectively.
  • The study used the Continuous Ranked Probability Score (CRPS) to evaluate the probabilistic forecasts of the hybrid models, with the GP-LSTM and GP-CNN models reporting lower CRPS values compared to the GP-RBM model.

Sources:

  • "Hybrid modeling approaches for predicting COVID-19 mortality: A comparative study across USA, France, and India," Results in Engineering, 2025, 26():105092.
  • Virus Weekly, "Research from Hyderabad Has Provided New Study Findings on COVID-19 (Hybrid modeling approaches for predicting COVID-19 mortality: A comparative study across USA, France, and India)," June 10, 2025, p 279.

Read more