Conceitos essenciais
Extending the Katz index to incorporate spatial and temporal patterns of fish movement significantly improves the prediction of farms susceptible to disease via live fish transfers.
Resumo
This study explores the use of Katz index-based models to enhance link prediction in live fish movement networks, with the goal of improving disease spread models in aquaculture. The authors developed several variants of the Katz index, including the Weighted Katz Index (WKI), Edge Weighted Katz Index (EWKI), and combined models (KIEWKI, WKIEWKI), to incorporate spatial and temporal factors.
The key findings are:
The EWKI model significantly outperformed the traditional Katz index (KI) and other variations, achieving high precision (0.988), recall (0.712), F1-score (0.827), and AUPR (0.970). This demonstrates the value of incorporating spatial distance between farms into the link prediction model.
The combined models (KIEWKI, WKIEWKI) approached, but could not surpass, the performance of the EWKI model. These models provide a balance between precision and recall, allowing for more flexible application in disease management scenarios.
The study highlights that the proximity of fish farms and the temporal dynamics of live fish movements are crucial factors in accurately predicting disease transmission pathways. Incorporating these spatial and temporal features into the Katz index-based models significantly improves their ability to identify potential links between farms.
The enhanced link prediction models can inform more accurate and responsive disease spread models, enabling better surveillance, biosecurity, and control measures in aquaculture networks.
Estatísticas
The dataset includes 16,946 live fish movements between 2,480 fish farms across England and Wales from 2010 to 2023.
Citações
"The EWKI model's performance demonstrates an innovative and flexible approach to tackling spatial challenges within network analysis."
"This study highlights the value of extending Katz index models to improve disease spread predictions in aquaculture networks."