Risk control model modeling method and device, enterprise risk assessment method and device and storage medium

Abstract

The invention relates to a risk control model modeling method. The invention discloses a device, computer equipment and a storage medium. The method comprises the following steps of obtaining a positive sample data set and a negative sample data set, Authentication datasets, inputting the verification data set, the positive sample data set and the negative sample data set into a to-be-trained machine learning model for training until training conditions are met, obtaining an enterprise scoring model, taking the verification data set, the positive sample data set and the negative sample data set as sample characteristics of an abnormal inspection scoring model, and obtaining an abnormal inspection scoring model; combining The enterprise scoring model and the abnormal inspection scoring model to obtain the risk control model, and combinding the enterprise scoring model with the abnormal inspection scoring model to perform enterprise risk scoring, so that the result error is reduced, andthe accuracy and the stability of the risk control model are further improved.

Description

technical field [0001] The present application relates to the technical field of the Internet, in particular to a risk control model modeling, enterprise risk assessment method, device, computer equipment and storage medium. Background technique [0002] With the continuous development of Internet technology, according to business needs, it is necessary to conduct corporate risk judgments for companies that intend to become cooperative relationships, so as to avoid unnecessary losses. The manual collection of enterprise-related information is converted to the use of Internet technology to collect data. According to user The empirical judgment of corporate risk is changed to a risk control model for corporate risk assessment. [0003] At present, risk control modeling generally uses traditional machine learning methods, such as using a logistic regression model to create a scorecard. This supervised learning method is learned based on negative samples of historical data, and...

AI Technical Summary

Problems solved by technology

Positive and negative samples need to be obtained through related acquisition methods, such as: according to the transaction process with customers, but before the transaction, the risk control system will decide whether to trade with the customer based on the risk judgment of the customer. When the customer is a high-risk customer, they will refuse to trade with the customer, because the rejected customers often have no labels, resulting in a large difference between the distribution of training samples and the distribution of real data, which leads to the accuracy and stability of the established risk control model Sex is not high

Images