A method for constructing an evaluation index of a label system and related devices

Abstract

This application discloses a method for constructing an evaluation index of a label system and a related device. On the one hand, the label cluster clustering results of multiple historical interaction samples are obtained; According to the first label of a historical interaction sample, multiple historical interaction samples are clustered to obtain the first prediction cluster clustering result. According to the similarity between the clustering results of the labeled clusters and the clustering results of the first predicted clusters, an evaluation index of the first labeling system is constructed. This evaluation index can reflect the similarity between the clustering results of the first predicted cluster and the clustering results of the labeled clusters. The higher the similarity, the more consistent the clustering results of the first predicted cluster are with the real situation, and the more reasonable the discrimination of the first label system is. , so the evaluation index of the first labeling system can more accurately measure the pros and cons of the constructed labeling system, and based on this evaluation index, a labeling system with a reasonable degree of discrimination can be constructed more accurately.

Description

technical field [0001] The present application relates to the field of data processing, in particular to a method for constructing an evaluation index of a label system and a related device. Background technique [0002] User Profile (User Profile) is a tagged and virtual user model abstracted based on real information such as user basic attributes, social attributes, behavioral attributes, psychological attributes, etc., in order to understand user needs and preferences through user portraits, and then provide users to provide better services. As an effective tool to outline target users, connect user appeals and design directions, user portraits have been widely used in various fields such as dialogue products, personalized recommendations, and precision marketing. [0003] In order to create a user portrait, it is necessary to define the portrait features of each dimension and the value of the portrait label of each dimension, that is, to construct the label system of th...

AI Technical Summary

Problems solved by technology

The current labeling system is either too coarse to distinguish, or too fine to make the labeling system too complex to be universal.

Images