User tag information determination method and device, equipment and storage medium

Abstract

The embodiment of the invention discloses a user tag information determination method and device, equipment and a storage medium. The method comprises the steps of obtaining target interaction behavior data of a target user; determining a target classification result corresponding to the target user according to a preset classification model and the target interaction behavior data; determining target label information of the target user according to the target classification result, wherein the preset classification model is obtained by training according to sample interaction behavior data corresponding to the sample user and standard label information in advance, and the standard label information is determined based on the sample interaction behavior data. Through the technical scheme of the embodiment of the invention, the accuracy of determining the user tag can be improved.

Description

technical field [0001] The embodiments of the present invention relate to Internet technologies, and in particular to a method, device, device and storage medium for determining user tag information. Background technique [0002] With the rapid development of Internet technology, it is often necessary to create user portraits in order to achieve refined operations and better provide users with targeted services, such as personalized item recommendations. User portraits can refer to the process of abstracting data into tags and using tags to concretize user images. [0003] The traditional method of user portraits is to manually label users, but with the increase in the number of users and the diversification of user behavior and needs, this traditional method is inefficient and difficult to guarantee the accuracy of labeling. In view of this, in the prior art, the classification model can be trained based on the user's historical behavior data and real labels, and the user ...

AI Technical Summary

Problems solved by technology

[0005] The existing classification model is based on the user's real physiological data to train and predict the user label results, but each user may play different social roles in different time periods, such as single role, husband role and parent role, etc., not only It is limited to the single role of the individual, so it is difficult to predict the accurate user label results only by using the user's real physiological data, which will also reduce the item recommendation effect

For example, if a user's real biological gender is male, but the user is in a relationship and often pays attention to female information, if the user's historical behavior data is marked as male based on the user's real biological gender and model training is performed, then the training The final classification model will also predict the historical behavior data as male. When the subsequent business party recommends male item information to the user based on the male label, the user pays more attention to female information, which will lead to recommendation The male item information for this user is ineffective, reducing the conversion rate of the item

Images