基于裁判文书挖掘的食品质量安全监管模式识别与绩效研究

食品质量安全监管是近年来中国社会最为关注的问题之一，我国为此从立法、执法上都加强了对食品质量安全的治理，但相关事件仍然屡有发生，食品质量安全监管政策的效力、执行能力和监管效果仍然有待提高。本项目的研究目标是解决如何从记录有政策措施、执法过程与处罚结果信息的大量裁判文书数据中挖掘食品质量安全事件特征，设计一套准确识别监管措施与模式和有效评估其绩效的方案的问题。通过融合食品质量安全事件的监管行为与语义特征数据，开发和应用基于张量分解的监管模式要素聚类模型和监管绩效预测模型，分别识别监管模式和评估事件的监管绩效，并分析大量事件中不同监管措施与模式产生的绩效，解决上述问题。本研究有助于从现行食品质量安全监管体制存在的问题切入，拓展监管体制研究的理论与方法，为我国持续推进食品质量安全监管体制改革提供理论与实践决策支持。

Food quality and safety supervision is one of the most concerned issues in China in recent years. Our country has strengthened the governance of food quality and safety through legislation and law enforcement, but related incidents still occur frequently, indicating that the effectiveness, the ability to implement and the supervision effects of food quality and safety supervision policies still need to be improved. The research goal of this project is to solve the problem: how to mine the characteristics of food quality and safety incidents from a large number of judgment documents that record information on policy measures, law enforcement processes and punishment results, and design a plan to accurately identify supervision measures and models and effectively evaluate its performance. The method to solve this problem is to integrate the supervision behavior and semantic feature data of food quality and safety events, develop and apply a clustering model and a performance prediction model based on tensor decomposition to identify the supervision pattern and evaluate the supervision performance of the events, and analyze the performance of different supervision measures and models in a large number of incidents. The research of this project is helpful to start from the problems existing in the current food quality and safety supervision system, expand the theory and method of supervision system research, and provide theoretical and practical decision support for the continuous promotion of food quality and safety supervision system reform in China.