Outbreaks of infectious diseases have caused tremendous human suffering and incalculable economic losses, and infectious diseases are a global public health problem that threatens human society.
Therefore, it is necessary to model the spatial and temporal distribution characteristics of infectious diseases, explore the transmission trend of infectious diseases, establish an infection early warning model and take corresponding preventive and control measures, which can make the prevention and control work more targeted and forward-looking.
Given the complex spatial correlation and temporal variation of infectious diseases, deep learning-based Spatio-temporal sequence prediction is widely employed because of its superior performance in capturing Spatio-temporal features.
However, current deep learning-based infectious disease prediction methods utilize an encoder-decoder structure that provides barely satisfactory accuracy due to a lack of understanding of infectious disease prevalence factors or deficiencies in capturing representative Spatio-temporal patterns.
In this paper, we develop the Transformer-Enhanced Spatio-Temporal Network (TEST-Net) which consists of a temporal location coding module and a Spatio-temporal feature fusion module for Infectious disease prediction.
Temporal information is input in TEST-Net by Temporal Location Encoding (TLE), and temporal and spatial correlation of sequences is extracted by a transformer-based attention network, and temporal features are fused with spatial features by a Spatio-temporal feature fusion network.
Compared with other state-of-the-art methods, qualitative and quantitative results show that TSET-Net has an excellent performance in modeling the spatial and temporal distribution characteristics of data and performs well in the accuracy of long-term prediction of infectious disease.