User default prediction method and device and electronic equipment

Abstract

The invention relates to a user default prediction method and device, electronic equipment and a computer readable medium. The method comprises the following steps: obtaining financial data of a user,wherein the financial data comprises a plurality of types of feature data; performing character conversion on the financial data to generate user data; and generating a default category and a defaultprobability of the user through the user data and a user default prediction model, wherein the user default prediction model is obtained by training a distributed gradient boosting decision tree model. According to the user default prediction method and device, the electronic equipment and the computer readable medium, the model training and calculation time can be saved, the grid search time ofa model is saved, and the user default prediction efficiency and accuracy are improved.

Description

technical field [0001] The present disclosure relates to the field of computer information processing, and in particular, to a user default prediction method, device, electronic equipment, and computer-readable medium. Background technique [0002] Usually, the machine learning model needs to learn positive samples and negative samples. The positive samples are the samples corresponding to the correctly classified categories. In principle, the negative samples can select any other samples that are not the correct category. The machine learning model establishes specific tasks based on positive and negative samples, and then trains the machine learning with specific data. After the training, a machine learning model suitable for a specific task is obtained. [0003] At present, most of the existing methods for user default risk analysis are based on decision tree methods, such as GBDT (Gradient Boosting Decision Tree), which are basically based on pre-sorted decision tree alg...

AI Technical Summary

Problems solved by technology

[0003] At present, most of the existing methods for user default risk analysis are based on decision tree methods, such as GBDT (Gradient Boosting Decision Tree), which are basically based on pre-sorted decision tree algorithms. The decision tree-like algorithm needs to save the eigenvalues ​​of the data, and also needs to save the result of feature sorting, such as the sorted index. In order to quickly calculate the split point in the future, the risk analysis model generated by this method needs to consume twice as much training data. Memory

Secondly, the risk analysis model generated by this method also has a large overhead in calculation time. When traversing each split point, it is necessary to calculate the split gain, which consumes a lot of money.

As more and more users are served on the financial service platform, and users have increasingly stringent requirements on response time, most of the current default risk analysis models cannot meet the timeliness requirements

Images