Presentation Information
[SS10-04]Forecasting Seasonal Infectious Disease Outbreaks using Statistical and Machine Learning Methods
*Hyojung Lee1 (1. Department of Statistics, Kyungpook National University (Korea))
Keywords:
Seasonal infectious diseases,Outbreaks
Norovirus is a primary viral cause of gastroenteritis in humans, exhibiting pronounced seasonal patterns characterized by a sharp increase primarily during winter months. With no available preventive vaccine, accurate prediction of outbreaks and the establishment of effective preventive strategies are critically needed. Influenza similarly represents a notable seasonal respiratory infectious disease, exhibiting high incidence rates from winter to early spring in Korea. This seasonal phenomenon is attributed to viral survival and transmission facilitated by low temperatures and dry conditions. Consequently, accurate prediction and proactive preparation for influenza outbreaks are essential for protecting public health and efficiently managing healthcare resources.We aim to improve long-term outbreak prediction accuracy by developing a hybrid predictive model that integrates statistical modeling and machine learning techniques. The approach involves combining meteorological data such as temperature and precipitation with epidemiological information. We evaluated and compared the predictive performance of various models, including the SARIMA model for statistical forecasting, the Long Short-Term Memory (LSTM) neural network for machine learning, and a hybrid model leveraging their respective strengths.By predicting patient numbers and analyzing outbreak periods for seasonal infectious diseases, the outcomes of this research are expected to significantly contribute to enhancing preparedness and developing more effective preventive strategies against seasonal infectious diseases.