认罪认罚案件量刑建议的大数据模型建构

The Big Data Model of Sentencing Proposals of Cases with Leniency on Admission of Guilt and Acceptance of Punishment

量刑建议精准化的关键在于提高量刑能力,定量与自由裁量并行的双轨制量刑方法是较优的选择。可运用数学量刑方法,以法官集体经验为基础,构建量刑模型,并结合案件具体情节对计算结果予以调整。根据"以刑制罪"理念,对犯罪行为具有相似性的犯罪,可构建通用量刑模型,而无需对每个特定罪名分别建模。对不同类型的犯罪,同一量刑情节的适用方式及对量刑结果的影响程度存在差异。调整模型计算结果时,需综合考虑犯罪类型、刑事政策、地理文化特征等具体情况,确保量刑公正。

The key to promoting the accuracy of sentencing proposals is to improve the ability of sentencing. The dual-track approach with both quantitative analysis and discretion is a method for sentencing. It’s necessary to establish a model for sentencing with mathematical methods based on the collective experience of judges and adjust the judgements according to the details of cases. On the basis of the idea“Decide the crime according to the penalty it deserves”, for crimes that have similarities to some extent, the establishment of a common model for sentencing is possible. A specialized model for every single crime is unnecessary. For different types of crimes, the application way of the same circumstance of sentencing and its effect on sentencing varies. To ensure the sentencing justice, specific conditions like the type of the crime, the criminal policies, the geographic and cultural features should be considered when judgements from the model are adjusted.