Regression model training method and device, electronic equipment and storage medium

Abstract

The invention provides a regression model training method and device, electronic equipment and a storage medium. The training method comprises the steps of obtaining features of labeled users and labeled label data in a labeled sample pool; training to obtain a regression model according to the features of the labeled users and the label data; obtaining a corresponding query strategy according to the regression model; determining a corresponding target unlabeled user according to the query strategy; obtaining features of a target unlabeled user and labeled label data; and adding the features of the target unlabeled user and the labeled label data to the labeled sample pool. Therefore, according to the method, the corresponding target unlabeled user is determined according to the query strategy, the features of the target unlabeled user and the labeled label data are added to the labeled sample pool, and training of the regression model is realized by adopting an active learning mode; and compared with the prior art in which model training is carried out based on a training sample obtained through random sampling, the generalization ability of the model is improved.

Description

technical field [0001] The present application relates to the technical field of deep learning, and in particular to a regression model training method, device, electronic equipment and storage medium. Background technique [0002] At present, the regression model has been widely used in fields such as user characterization and enterprise decision-making. For example, the regression model can describe the user's income information based on the user's basic information, consumption records, and consumption preferences (for example, favorite products); or, the regression model can be based on the company's sales information, industry information, business status and other information to predict the registered capital of the enterprise. However, most of the regression model training methods in the related art are based on completely random sampling or stratified sampling to obtain training samples, which has poor flexibility. Contents of the invention [0003] This applicat...

AI Technical Summary

Problems solved by technology

However, most of the regression model training methods in related technologies are based on completely random sampling or stratified sampling to obtain training samples, which is less flexible.

Images