MPSTAN: Metapopulation-based Spatio-Temporal Attention Network for Epidemic Forecasting

TL;DR

The paper introduces MPSTAN, a novel spatio-temporal neural network that incorporates multi-patch epidemiological knowledge and inter-patch interactions to improve the accuracy and stability of epidemic forecasting.

Contribution

It proposes a hybrid model that integrates domain knowledge into both model construction and loss functions, enhancing epidemic prediction accuracy.

Findings

Outperforms baseline models in accuracy and stability.

Using domain knowledge in both model and loss improves forecasting.

Effective incorporation of multi-patch epidemiological knowledge enhances results.

Abstract

Accurate epidemic forecasting plays a vital role for governments in developing effective prevention measures for suppressing epidemics. Most of the present spatio-temporal models cannot provide a general framework for stable, and accurate forecasting of epidemics with diverse evolution trends. Incorporating epidemiological domain knowledge ranging from single-patch to multi-patch into neural networks is expected to improve forecasting accuracy. However, relying solely on single-patch knowledge neglects inter-patch interactions, while constructing multi-patch knowledge is challenging without population mobility data. To address the aforementioned problems, we propose a novel hybrid model called Metapopulation-based Spatio-Temporal Attention Network (MPSTAN). This model aims to improve the accuracy of epidemic forecasting by incorporating multi-patch epidemiological knowledge into a spatio-temporal model and adaptively defining inter-patch interactions. Moreover, we incorporate inter-patch epidemiological knowledge into both the model construction and loss function to help the model learn epidemic transmission dynamics. Extensive experiments conducted on two representative datasets with different epidemiological evolution trends demonstrate that our proposed model outperforms the baselines and provides more accurate and stable short- and long-term forecasting. We confirm the effectiveness of domain knowledge in the learning model and investigate the impact of different ways of integrating domain knowledge on forecasting. We observe that using domain knowledge in both model construction and loss functions leads to more efficient forecasting, and selecting appropriate domain knowledge can improve accuracy further.