Hybrid Prediction Model Based on Decomposed and Synthesized COVID-19 Cumulative Confirmed Data

Since 2020, COVID-19 has repeatedly arisen around the world, which had a significant impact on global economy and culture. The prediction of epidemic will help to deal with current similar risks that may arise in future. So, this paper proposes hybrid model based particle swarm optimization variational mode decomposition (PSO-VMD), Long Short-Term Memory Network (LSTM) AdaBoost algorithm. To address issue determining optimal number modes K penalty factor (α) (VMD), an adaptive value for (PSO) is proposed. Specifically, weighted average sample entropy relevant coefficients utilized determine value. First, data are decomposed into multiple modal components, known as intrinsic functions (IMFs), using PSO-VMD. These along policy-based factors, integrated form multivariate forecast dataset. Next, each IMF predicted AdaBoost-LSTM. Finally, results all components reconstructed obtain final result. Our proposed method validated by cumulative confirmed Hubei Hebei provinces. case confirmation data, coefficient determination (R2) mixed increased compared control model, mean absolute error (MAE) root square (RMSE) decreased. experimental demonstrate VMD–AdaBoost–LSTM achieves highest accuracy, thereby offering new approach prediction.