基于执行文书的大数据法律监督

Big Data Legal Supervision Based on Enforcement Documents

在大力推行执行信息共享与公开的现实背景下,海量执行文书为检察机关甄别批量异常案件提供了资源保障,可用以挖掘执行过程及诉讼、仲裁等活动中普遍性违法的线索及证据。基于执行文书的批量异常案件甄别模型以“批量”特征为关键,构成了一种契合检察监督需求的理论预设与数据驱动相结合的分析方法;可进一步提炼以“不法分子”为学习对象,以“数量特征+异常特征”为模型开发逻辑,以规则驱动为主要分析模式三大要素。以民间借贷领域三十万执行文书对模型进行验证,表明批量执行文书隐含着“合法但不合理型”和“明显违法型”两类异常问题。结合上述分析模型及其结果,检察机关可以配套分流方案与协作机制,以大数据法律监督为诉源治理提供检察智慧。

Under the current background of vigorous implementation of enforcement information sharing and disclosure,the massive volume of enforcement documents provides resources for procuratorial authorities to identify batch of abnormal cases,which can be used to uncover clues and evidence of universal violations in the enforcement process,litigation and arbitration activities.The key to the model of identifying abnormal cases based on enforcement documents is the“batch”feature,which is a theoretical predetermined and data-driven analysis method to meet the needs of procuratorial supervision.It can be summarized into three major elements:taking“violator”as the learning object,taking“quantitative features+abnormal features”as the logic of model development,and taking rule-driven as the main analysis mode.The model is validated with 300,000 enforcement documents in the field of private lending,showing that the batch enforcement documents imply two types of anomalies:“legal but unreasonable”and“clearly illegal”.Combined with the above analysis model and its results,we should support the triage program and collaboration mechanism,in order to provide strong support for big data legal supervision,and offer prosecution wisdom for litigation source governance.