Master Assignment
Enhancing COVID-19 Infection Network Inference
Type: Master
Period: TBD
Student: (Unassigned)
If you are interested, please contact :
Background
In recent research, Prasse and Van Mieghem (2020) introduced the Network Inference-based Prediction Algorithm (NIPA), a machine-learning method for constructing COVID-19 transmission networks from daily infection data across a country's regions. Central to NIPA's efficacy is selecting the right regularization parameter, which encourages the algorithm to prefer simpler -sparser- network solutions. However, the current approach to this selection involves testing a set number of values within a specific range, a somewhat arbitrary process. This research proposal aims to refine the selection process of the regularization parameter by adopting an optimization-based search approach. Utilizing publicly available daily COVID-19 data, the study seeks to enhance NIPA's precision and practical utility by exploiting metaheuristics like Simulated Annealing, Tabu Search, and Genetic Algorithms. Improved parameter selection could lead to more accurate predictions of COVID-19 spread, thereby informing more effective public health strategies and interventions. This research not only contributes to the academic field by optimizing an existing algorithm but also has the potential to offer tangible benefits in managing and understanding the COVID-19 pandemic dynamics.
References
B. Prasse and P. Van Mieghem, “Predicting Dynamics on Networks Hardly Depends on the Topology,” arXiv [physics.soc-ph], May 29, 2020. [Online]. Available: http://arxiv.org/abs/2005.14575
M. A. Achterberg, B. Prasse, L. Ma, S. Trajanovski, M. Kitsak, and P. Van Mieghem, “Comparing the accuracy of several network-based COVID-19 prediction algorithms,” Int. J. Forecast., vol. 38, no. 2, pp. 489–504, Apr. 2022.