Risk control model construction method and device based on multi-objective optimization and electronic equipment

Abstract

The invention relates to the technical field of computers, in particular to a risk control model construction method and device based on multi-objective optimization and electronic equipment, and the method comprises the steps: obtaining a business sample data set, and setting a plurality of corresponding labels for a feature group corresponding to each sample in the business sample data set; preprocessing the service sample data set according to the label to obtain a service sample training set; constructing an initial risk control model, wherein the initial risk control model internally comprises a plurality of cascaded logistic regression algorithms; and performing multi-target joint training on the initial risk control model according to the business sample training set to obtain a final risk control model. According to the method and device, excellent model indexes and business effects can still be obtained under the condition that model interpretability is guaranteed, the problem of sample deviation is greatly relieved due to the fact that multi-objective optimization joint training is used, and the constructed risk control model is good in effect, stable in performance and high in generalization ability after being actually online.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a method, device and electronic equipment for constructing a risk control model based on multi-objective optimization. Background technique [0002] For the risk control system of online consumer loans on the Internet, the current mainstream practice in the industry is to use a combination of manual strategies and machine learning models to build it. For manual strategies, because the strategy rules are generally relatively simple, it is easy to be broken by the black industry chain and the anti-fraud team through repeated trials and errors and collisions. For the machine learning risk control model, the current mainstream in the industry uses three types of algorithms based on the logistic regression algorithm (Logistic Regression), based on the decision tree algorithm, and based on the deep learning algorithm to build the model. [0003] The risk control model constr...

AI Technical Summary

Problems solved by technology

Artificial strategies are easy to break. Although the logistic regression algorithm has strong model interpretability, the business indicators are poor. Although complex models such as decision number algorithms and deep learning algorithms are effective, the interpretability of the model is poor, or even completely unexplainable. , which makes it very troublesome to face internal control and internal audits and inquiries from national regulatory agencies

Images