Network analysis of tuberculosis spread with forecasting and technique for order preference by similarity to ideal solution

Tuberculosis (TB), caused by Mycobacterium tuberculosis, remains a major health challenge, particularly in low income and middle-income countries. This study focuses on the Western province of Sri Lanka, which has high population density, trade, and high human mobility, all contributing to TB transmission. This research aims to strengthen TB control strategies through several key objectives. The TB contact network model was developed and analyzed using graph theoretical measurements such as degree centrality, closeness centrality and betweenness centrality, to identify influential nodes, and structural metrics such as modularity and clustering coefficient to understand transmission patterns. The TOPSIS (technique for order preference by similarity to ideal solution) method was applied to rank nodes based on their potential to spread TB, improving targeted intervention strategies. The novelty of this research is to apply graph theoretical measurements together with the TOPSIS method to identify the highly influential nodes and control the spread of disease by isolating those nodes from the community. Epidemic data extracted from the National Programme for Tuberculosis Control and Chest Diseases (NPTCCD) annual reports (2012–2023) revealed higher projected TB incidences among males and individuals belong to the age group 45–54. The synthetic dataset constructed using percentage-based forecasting data underwent 5-fold cross-validation, and the random forest model achieved an accuracy of 81.5%, indicating strong reliability of the forecasting model. Network analysis indicated the sparse local connections (clustering coefficient: 0.002) and strong community structures (modularity: 0.856). Comparative analysis before and after applying the TOPSIS method demonstrated the improved identification of influential spreaders. Sensitivity analysis showed that the model remained stable under small variations in data and weights, confirming the robustness of the findings. The outcome of this research provides guidance to health authorities for allocating resources, prioritizing high risk groups, initiating strengthening programmes, and designing more effective community-based interventions to strengthening TB control strategies in the Western province.