金融信贷风险管理大数据

金融信贷风险管理采集司法机关、行政机关、仲裁机构等公开的许可、处罚数据(不含个人数据)。将采集到数据通过DNN深度神经网络，从非结构化的文本中抽取命名实体得到我们需要的字段，然后对字段内容进行清洗，主要对字段值中的异常值、缺失值、无用值等进行标准化处理，之后通过数据分类算法按照机关、来源、事由、结果进行分级分类，最后依据排重规则生成数据的唯一主键进行消除处理。

This dataset for financial credit risk management collects public license and penalty data (excluding personal data) from judicial authorities, administrative agencies, arbitration commissions and other relevant institutions. The collected data is processed via DNN (Deep Neural Network) to extract named entities from unstructured texts to obtain the required fields. Subsequently, field content cleaning is performed, mainly standardizing abnormal values, missing values, invalid values and other problematic entries within field values. Next, a data classification algorithm is adopted to carry out hierarchical classification based on the authority, source, cause and outcome of the data. Finally, deduplication processing is completed by generating a unique primary key for each data record in accordance with the deduplication rules.

搜集汇总

数据集介绍

特点

该数据集名为'金融信贷风险管理大数据'，由北京全民普惠信用管理有限公司登记，属于信息传输、软件和信息技术服务业。数据来源于企业，每日更新，包含999,999条记录，主要用于金融信贷风险管理。数据通过DNN深度神经网络处理，进行实体抽取、清洗和分类。