基于分层学习的易混淆法条预测

目前针对法条预测的相关研究大都采用文本分类的思想,但模型构建过程都未考虑不同法条之间的从属关系或相似程度,因此对于易混淆法条预测效果普遍较差。针对现有方法在易混淆法条预测中存在的不足,提出基于分层学习的易混淆法条预测方法。将法条分为易区分法条和易混淆法条,按法条内容将易混淆法条组合为不同易混淆法条集并分别训练易混淆法条集预测模型,运用分层学习完成易混淆法条预测。在刑事案件的数据上进行实验,实验结果表明,该模型能较好解决易混淆法条预测问题,提高法条预测准确率。

Confusing law prediction based on hierarchical learning

At present,most of the related researches on the prediction of the law adopt the idea of text classification,however,the model construction process fails to consider the subordination or similarity between different laws,so the prediction effect on the confusing law is generally poor.Aiming at the shortcomings of existing methods in the prediction of confusing law,a confusing law prediction method based on hierarchical learning was proposed.The law was divided into two parts,namely distinguish law and confusing law.According to the content of the law,the confusing law was combined into different confusing law sets and the confusing law set prediction model was separately trained,and the hierarchical learning was used to complete the confusing law prediction.Experiments were carried out on the data of criminal cases.Experimental results show that the proposed model can better solve the problem of confusing law prediction and improve the accuracy of law prediction.