Research and Application of Time Series Forecasting Methods in Disease Surveillance

Abstract

Time series forecasting plays a pivotal role in public health, particularly within disease surveillance. Theoretically, it facilitates the in-depth analysis of disease patterns and progression; practically, it supports health authorities in optimizing resource allocation, devising prevention strategies, and mitigating transmission risks. Recently, rapid advancements in information technology have accelerated the development of this field, leading to the continuous emergence of novel methods. This paper provides a comprehensive overview of time series forecasting methods, categorizing them into traditional statistical models and deep learning approaches, while strictly evaluating their respective strengths and limitations. Furthermore, it explores innovative applications involving statistical models, machine learning, and deep learning, highlighting the potential of these emerging technologies. Bridging theory and practice, this systematic review aims to establish an analytical framework, delineate the field's developmental trajectory, and offer valuable insights for future research in disease prediction.