Risk control model construction method and device, storage medium and terminal

Abstract

The embodiment of the invention discloses a risk control model construction method and device, a storage medium and a terminal. The method comprises the steps of obtaining historical data of at leastone user in a preset time period; extracting at least one type of first feature information related to the historical data; for each type of first feature information, sorting the first feature information in the historical data according to a preset sorting rule to generate a first feature information sequence; determining a first arrangement quantile of the first feature information in the firstfeature information sequence; and training a preset machine learning model based on the first arrangement quantile to generate a risk control model. By adoption of the technical scheme, the corresponding feature information is replaced with the arrangement quantiles of the feature information in the historical data; and model training is performed based on the arrangement quantiles of the featureinformation to generate a risk control model, so that the constructed risk control model can achieve a dynamic self-adaptive effect on the change of the input data, and the attenuation speed of the risk control model along with time is effectively reduced.

Description

technical field [0001] The embodiments of the present invention relate to the technical field of data processing, and in particular to a method, device, storage medium, and terminal for constructing a risk control model. Background technique [0002] In the field of risk control, due to the influence of the macroeconomic environment and company strategies, the target user groups and user characteristics of the risk control model have been changing, making the effect of the built risk control model gradually decrease or become invalid over time. Faced with this situation, it is usually a period of time to use new data to fine-tune the original model to meet business requirements. [0003] However, the scheme of adjusting the original risk control model through new data will have the following two problems: 1. The modeling of recent new data will cause the new model to lose the information of the original model, and the distribution of output scores will generally be different...

AI Technical Summary

Problems solved by technology

[0003] However, the scheme of adjusting the original risk control model through new data will have the following two problems: 1. The modeling of recent new data will cause the new model to lose the information of the original model, and the distribution of output scores will generally be different from that of the original model. The difference is quite large, and relevant personnel need to do detailed testing and adjustments on the use strategy; 2. Since the supervised model in the field of risk control generally uses the number of overdue days of 3 months or 6 months as the modeling target, this makes the latest proposal Model data need to have a length of time of 3 months or 6 months, and the variability of risk control data can often change greatly within this period (3 months or 6 months), making this risk Controlling model adjustments have little effect

Images