A mpox deep learning prediction framework incorporating a two-strain SIRD epidemic model focusing on the cases of the Democratic Republic of the Congo A mpox deep learning prediction

• Published in: International journal of dynamics and control Vol. 13; no. 9
• Main Authors: Adak, Sayani, Pal, Sankar K.
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 02.09.2025
• Subjects: Complexity; Control; Control and Systems Theory; Dynamical Systems; Engineering; Vibration; Global sensitivity analysis; mpox modeling; mpox variants; mpox data of the DRC; Deep learning architecture
• ISSN: 2195-268X; 2195-2698
• DOI: 10.1007/s40435-025-01837-w

Historically endemic to Africa, mpox (formerly known as monkeypox) experienced its greatest outbreak in 2022, spreading to numerous parts of the world and posing a concern to public health. The WHO Director-General proclaimed on August 14, 2024, that the rise of mpox cases in the Democratic Republic of the Congo (DRC) and its spread to nearby nations constitute a Public Health Emergency of International Concern (PHEIC) as per the International Health Regulations (2005). Our present research focuses on the clade Ia and clade Ib outbreaks occurring in the Democratic Republic of the Congo. We formulate a two-strain susceptible–infected–recovered–deceased (TSSIRD) model using a system of ordinary differential equations. We predict the parameters derived from the TSSIRD model incorporating deep learning (DL) architectures such as LSTM, BiLSTM, Stacked LSTM, ConvLSTM, and ConvBiLSTM. The resulting TSSIRD-DL model is used for single-day prediction of the number of infected for both clades. The experimental results show that the TSSIRD-DL models predict significantly well in comparison with the pure deep learning models where the data are taken from January 1, 2023, to December 29, 2024. This approach can be fruitful for the single-day prediction of the infected population when only small historical data sets are available. Additionally, we have performed a global sensitivity analysis on the model parameters to understand their effects on basic reproduction numbers and infected individuals (for both clades). As a result, the significant parameters for the spread of the disease are found to be the disease transmission rate and the rate of recovery.