Safety monitoring data for agricultural products are currently limited as most tested indicators of the monitored samples are recorded as “undetected”, presenting a major research obstacle as researchers must still attempt to extract sufficient information. To acquire reference information for decision making and accurately quantify risks, researchers of agricultural product safety evaluation protocols need effective methods to combine the heterogeneous and sparse monitoring data from multiple monitoring projects.

Risk management is crucial to the safety of agricultural products. Focusing on the requirements of safety risk management of agricultural products, a product category and limited quantity-based strategy of data fusion was proposed to calculate the risk entropy of the “product + index” combination. Subsequently, using the theory and method of comprehensive evaluation, a macroscopic quantitative evaluation of agricultural products based on risk entropy was conducted.

According to a case analysis of the safety of pesticide residues in vegetables, risk entropy can effectively extract risk information in the fused data of risk monitoring and be used in quantitative analysis. Moreover, the risk-entropy-based safety evaluation can provide complete information about risk distribution, as well as the division of risk threshold and risk ranking. It was further found that the risk entropy can accurately identify potential risk, thus effectively avoiding overestimated or underestimated risks in qualitative evaluation.

The findings of this study indicate that the proposed risk-entropy-based data fusion and quantitative evaluation are technically feasible. This quantitative evaluation method has been shown to improve the performance of qualitative evaluation in risk identification and description. In conclusion, the method increases the abundance of accurate references for safety risk management of agricultural products, thus promoting the progress of agricultural product safety risk management.