The impact of scholarship inflows on achieving food security: what can Bayesian networks tell us?

This study uses Bayesian networks besides the regression model to analyse the relationship between scholarship inflows and food security in 46 of Belt and Road Initiative countries. Our main contributions are: 1–Analysing the relationship between scholarship inflows and food security. 2–Using Bayesian networks to analyse this relationship. 3–Comparing Bayesian networks results with regression model results. The regression model resulted that scholarship inflows have a significant positive effect on food security with small coefficient value. On the other hand, Bayesian networks showed that food security conditionally depends on scholarship inflows given the percentage of agricultural value add of GDP. In addition, Bayesian networks had higher prediction accuracy than the regression model. Whereas the constraint-based approach of network structure learning showed the highest prediction power, the information theory measures of network quality, including Entropy and Mutual Information, revealed better performance than Bayesian measures. Finally, we concluded that using Bayesian networks beside linear models could enhance our results.