Model label labeling method and device based on support vector machine and medium

Abstract

The embodiment of the invention discloses a model label labeling method and device based on a support vector machine and a medium, and the method comprises the steps of building an index system of a credit model according to the credit data of an enterprise, and determining the index data; classifying preset index data in the index data, and labeling labels; according to the labeled preset index data, training a pre-constructed sample label labeling model to obtain a sample label labeling model meeting requirements, and determining a corresponding relationship between index features of the preset index data and labeled labels, the sample label labeling model being a support vector machine model; analyzing the index characteristics of the unlabeled index data in the index data through the sample label labeling model meeting the requirements, and classifying and labeling the unlabeled index data according to the corresponding relationship between the index characteristics of the preset index data and the labeled labels, so as to train a credit model according to the labeled index data, and obtaining a credit model meeting the requirements.

Description

technical field [0001] This specification relates to the field of computer technology, in particular to a model labeling method, device and medium based on a support vector machine. Background technique [0002] The development and application of Internet technology has penetrated into various industries, especially the financial industry. The traditional financial industry is facing the severe test of big data, the Internet, and the imbalance and asymmetry of user experience. The use of big data risk control technology has become a big data application focus that the industry pays the most attention to. [0003] In the current Internet financial credit system, the credit scoring model is usually used to evaluate the credit of enterprises. The common method of credit scoring model is logistic regression. When using logistic regression to build a credit scoring model, a large number of labels containing However, in the actual application environment, in many cases, a large a...

AI Technical Summary

Problems solved by technology

[0004] One or more embodiments of this specification provide a model labeling method, device, and medium based on support vector machines, which are used to solve the following technical problems: In actual application environments, most of the labeled data cannot be obtained, resulting in The training samples of the credit model are insufficient, and a credit scoring model that meets the requirements cannot be trained

Images