Object clustering method, and model training method and device

Abstract

The embodiment of the invention discloses an object clustering method, and a model training method and device. The objective of the invention is to solve the problem that human intervention factors in the clustering division process influences the accuracy of the final clustering division in the prior art. The object clustering method comprises: determining a Euclidean distance of any two objects in an object set according to a preset feature value corresponding to each object in the object set to be clustered; performing distribution of objects in the object set in a three-dimensional space according to the Euclidean distance; determining a first core object of the number of objects in a neighborhood with a preset radius being not smaller than a preset value based on the distribution of the objects in the object set in the three-dimensional space; determining a second core object of the number of objects in the neighborhood with the preset radius being not smaller than the preset value and located in the neighborhood with the preset radius of the first core object; and belonging the first core object and the second core object to the same object cluster.

Description

technical field [0001] The present application relates to computer technology, in particular to an object grouping method, a model training method and a device. Background technique [0002] At present, machine learning (Machine Learning, ML) is applied in various fields of artificial intelligence. For example: machine learning is performed by extracting user credit data to obtain a credit scoring model. [0003] Taking the application of credit scoring as an example, the credit data of all users is generally used for machine learning to obtain a unified credit scoring model, and the credit evaluation of all users is completed in this unified credit scoring model. Practice has proved that due to the existence of different group characteristics or population distribution in user groups, it is often difficult for the above-mentioned unified credit scoring model to achieve satisfactory credit evaluation results. Therefore, it has become an important link in the current credit...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the embodiment of the present application is to provide an object grouping method, a model training method and a device to solve the problem in the prior art that the accuracy of the final group division is affected by factors of human intervention in the group division process

Images