Evaluating socioeconomic influences on Nipah virus vaccination decisions in Bangladesh through machine learning.

Nipah virus (NiV) is emerging as a noteworthy public health threat in Bangladesh, with propagation transpiring across accidental person-to-person contact and proximity with contaminated food or animals. Addressing the socio-demographic elements impacting the awareness of NiV is essential for productive public health initiatives. This study explores the impact of socio-demographic characteristics on individuals' willingness to get vaccinated (WGTV) for the Nipah virus during an outbreak in Bangladesh and to develop optimal control strategies using machine learning techniques. A cross-sectional survey in Bangladesh examined socio-demographic factors, health conditions, and common misconceptions about vaccination, and statistical analysis was performed using chi-square tests to identify significant associations between demographic variables and vaccination willingness. Machine learning algorithms, including Random Forest (RF), Decision Tree (DT), and Extreme Gradient Boosting (XGBoost), were employed to predict vaccination willingness and determine feature importance. SHAP (Shapley Additive Explanations) analysis further validated the significance of these features in two distinct scenarios. We analyzed two distinct scenarios in this study. In the first scenario, we utilized all five categories for classification for WGTV, achieving a maximum accuracy of 60.83% with the Random Forest (RF) model. In the second scenario, to address the issue of class imbalance, we transformed the study into a binary classification problem, which significantly improved the model's performance, yielding the highest accuracy of 85% with the RF model. Notably, the model outputs were influenced by demographic parameters such as age and daily interactions, but their influence was mild compared to the predominant influence of socioeconomic characteristics. Analysis validated key predictors, highlighting socioeconomic factors like education and income as the most influential, while demographic factors such as age and gender had a milder effect. The findings underscore the importance of addressing socioeconomic disparities in public health efforts to boost immunization rates, offering valuable insights for policymakers to design targeted interventions and improve health outcomes.