Application of Bayesian network in water quality risk analysis and pollution reduction decision making from small data

As an effectual tool for risk analysis, Bayesian networks have a great capability of quantifying risk, and a great flexibility to adapt to little information. In this paper, we first establish a Conditional Linear Gaussian Bayesian network representing the water quality of Section Yidong Bridge, Dongyang River, China, using small data. Then we evaluate the risk of pollutant concentration exceeding standards. Finally, based on the risks calculated, we recommend pollution reduction decisions, which can be considered as a reference proposal for administrators under the condition of small data.