Source detection in epidemic dynamics on hypergraphs using a dynamic message passing algorithm

Area of Science:

• Epidemiology
• Network Science
• Computational Biology

Background:

• Source detection is vital for managing infectious disease outbreaks and implementing control measures.
• Traditional methods often rely on pairwise networks, neglecting complex group interactions.
• Hypergraph representations are increasingly recognized for capturing group-based transmission patterns.

Purpose of the Study:

• To develop a novel message passing algorithm for accurate source detection in infectious diseases.
• To incorporate group interaction information, represented by hypergraphs, into source detection models.
• To evaluate the performance of the proposed algorithm against existing methods.

Main Methods:

• A message passing algorithm, termed HDMPN (Hypergraph-based Dynamic Message Passing Network), was developed.
• The algorithm modifies likelihood maximization by utilizing the proportion of infectious neighbors within hyperedges.
• Stochastic susceptible-infectious dynamics with correlated infections within hyperedges were modeled.

Main Results:

• The HDMPN algorithm demonstrated superior performance in source detection compared to benchmarks.
• Incorporating hyperedge information significantly improved the accuracy of identifying disease origins.
• The proposed method effectively captures correlated infection events within group interactions.

Conclusions:

• The HDMPN algorithm offers a more accurate approach to source detection by accounting for group transmission dynamics.
• Hypergraph representations are crucial for understanding and modeling complex epidemic propagation.
• This work advances the field of infectious disease modeling and source detection.