Risk control model application methods and systems based on concurrence of multiple machine learning models

Abstract

The method discloses a risk control model application methods and systems based on a concurrence of multiple machine learning models. An offline risk control model application method thereof includes:collecting application customer group information; extracting customer portrait data from the above-mentioned information; processing the above-mentioned portrait data, and calculating risk-related indexes; respectively utilizing different feature selection models to process the above-mentioned risk-related indexes to obtain corresponding single-model feature orders; carrying our comprehensive sorting on the obtained corresponding single-model feature orders; and respectively utilizing multiple machine learning models to carry out modeling processing on comprehensively sorted features, evaluating a running effect of each machine learning model, and screening out the plurality of machine learning models with higher ranks. The plurality of machine learning models are put into online runningto process online application flow, and processing results are used as decision bases of whether credit extension is carried out. The flow is allocated in online running according to the rank of eachmachine learning model. By adopting the methods, efficiency of risk evaluation can be improved, and a result is more reliable.

Description

technical field [0001] The invention relates to a method and system for risk assessment. Background technique [0002] Credit risk control is of great significance to the healthy and sustainable development of the economy. Credit risk assessment is an important means of credit risk control. Traditional credit risk assessment relies too much on the central bank's credit reporting system, but the credit reporting system's credit reporting data has shortcomings in terms of timeliness, comprehensiveness, and hierarchy, and cannot fully reflect the real information of customers. In addition, the common model in risk assessment is the expert model, which relies too much on the personal experience of experts, is highly subjective and arbitrary, and the assessment results are not objective and accurate enough, and the efficiency is low. Contents of the invention [0003] The present invention provides a wind control model application method and system based on multiple machine l...

AI Technical Summary

Problems solved by technology

Traditional credit risk assessment relies too much on the central bank's credit reporting system, but the credit reporting data of the credit reporting system has shortcomings in terms of timeliness, comprehensiveness, and hierarchy, and cannot fully reflect the real information of customers

In addition, the common model in risk assessment is the expert model, which relies too much on the personal experience of experts, is highly subjective and arbitrary, and the assessment results are not objective and accurate enough, and the efficiency is low

Images