Method and system for analyzing and predicting bank personal loan overdue by utilizing characteristic trend

Abstract

The invention discloses a method and a system for analyzing and predicting bank personal loan overdue by utilizing a characteristic trend, and the method comprises the following steps: obtaining to-be-processed customer data, and marking a positive sample; performing characteristic box separation; calculating a positive sample proportion; establishing a trend change relationship according to the proportion of the sub-boxes and the positive samples, performing similarity calculation on the trend change relationship, and screening out features of which the similarity is greater than a first threshold value; performing feature aggregation processing on the screened features to obtain new features; segmenting the client data according to a certain proportion, repeating the above steps, comparing the change conditions of the target values of the same features along with the change of the feature values under different data volumes, obtaining feature noise according to the change conditions,manually selecting the features of which the feature noise is higher than a noise threshold value, and taking the features lower than the noise threshold and the manually selected features as modeling features; and constructing a bank personal loan overdue prediction model by using the modeling features, and predicting loan overdue according to the model to obtain a prediction result.

Description

technical field [0001] The invention relates to the technical field of bank data processing, in particular to a method and system for predicting overdue bank personal loans by using feature trend analysis. Background technique [0002] Bank personal loan overdue prediction has important guiding significance for bank lending. However, in the actual modeling process, a lot of customer information is involved, and the customer's capital transactions in the bank involve many dimensions; if such data is directly modeled It is often difficult to obtain good results. [0003] In the existing technology, the prediction of personal loan overdue using feature engineering is often based on experience; when dealing with multi-dimensional features that customers have more capital transactions with banks, these features cannot be further analyzed by relying on experience. [0004] Therefore, there is an urgent need for a technical solution that can explore the trend of characteristic cha...

AI Technical Summary

Problems solved by technology

[0002] Bank personal loan overdue prediction has important guiding significance for bank lending. However, in the actual modeling process, a lot of customer information is involved, and the customer's capital transactions in the bank involve many dimensions; if such data is directly modeled often difficult to achieve good results

[0003] In the existing technology, using feature engineering to predict personal loan overdue is often based on experience; when dealing with multi-dimensional features that customers have more capital transactions with banks, these features cannot be further analyzed by relying on experience

Images