Research on the construction method of corporate credit risk portrait based on deep learning
To better characterize corporate credit risk characteristics,this study proposes a method for constructing corporate credit risk portraits based on data distribution and deep learning.First,an enterprise credit risk label system is constructed based on the enterprise credit risk model,focusing on characterizing the group comparison characteristics in the spatial dimension and the individual change characteristics in the time dimension.Then,for statistical labels,feature labels are generated according to specific rule definitions;for mining/statistical labels,deep learning technologies such as named entity recognition and emotion clas-sification are used to construct feature labels based on specific tasks.Corporate credit risk portraits can intuitively reveal corporate credit risk characteristics.Deep learning technology has effectively expanded corporate characteristics characterization capabilities and better served credit regulation and smart regulation.
deep learningcorporate profiledata distributioncredit risks
万方数据
维普
