Comparison on Machine Learning Methods for Infectious Diseases Prediction

The primary focus of this study is to explore the predictive effectiveness of various machine learning and time series models. By leveraging three years of pandemic data from China, the research aims to identify the optimal predictive models for different infectious disease patterns, thereby contributing significantly to future pandemic prognosis and providing rigorous validation for pandemic prevention and control measures. This comprehensive study reviews previous research and selects the most representative and validated predictive models. Most of these models have been used to predict infectious disease include the Seasonal Autoregressive Integrated Moving Average (SARIMA), Exponential Smoothing State Space Model (ETS), Long Short-Term Memory (LSTM), Hybrid Models, Trigonometric, Box-Cox transformation. By incorporating these advanced predictive models, the study aims to improve research efficiency and accuracy in forecasting infectious disease trends. The ultimate goal is to provide robust tools and methodologies that can be utilized for effective pandemic management, helping policymakers and health professionals to make informed decisions and implement timely interventions.