Forecasting Global Monkeypox Infections Using LSTM: A Non-Stationary Time Series Analysis

This study leverages the capabilities of Long Short-Term Memory (LSTM) models in forecasting global Monkeypox infections, thereby demonstrating the significant potential of advanced machine learning techniques in epidemiological forecasting. Our LSTM model effectively navigates the challenges posed by non-stationary time-series data, a common issue in epidemiological studies. It successfully captures the underlying patterns in the data, producing reliable forecasts. The model’s performance was evaluated using several metrics, including RMSE, MSE, MAE, and R ² , all of which pointed to its robust and satisfactory predictive capabilities. Our findings underscore the significant role LSTM models can play in informing the development of timely and effective disease control and prevention strategies. They thereby contribute to enhancing public health responses to emerging infectious diseases such as Monkeypox. However, despite the promising results, the study highlights the ongoing challenge of enhancing the interpretability of LSTM models, an area that warrants further research. As a future direction, efforts should focus on refining LSTM models to bolster their interpretability, ensuring their broader adoption and utility in public health practice.