Corpus labeling method, device, server and storage medium

Abstract

The embodiments of the present invention relate to the technical field of information processing, and in particular to a corpus tagging method, device, server and storage medium. The corpus labeling method includes: obtaining an even-numbered manual labeling result for the initial corpus, and a model labeling result for the initial corpus; wherein, the model labeling result for the initial corpus is obtained according to a preset labeling model, and the preset The labeling model is obtained according to several initial corpus trainings that have been manually labeled; among all the labeling results including the manual labeling results and the model labeling results, the only labeling result that meets the preset conditions is obtained as the initial corpus The final labeling result. By adopting the embodiment of the present invention, high-quality corpus labeling results of the initial corpus can be obtained, and the influence of a single labeler on the quality of corpus labeling can be reduced.

Description

technical field [0001] The embodiments of the present invention relate to the technical field of information processing, and in particular to a corpus tagging method, device, server and storage medium. Background technique [0002] Natural language processing refers to the computer receiving input in the form of natural language, and internally processes and calculates the natural language through user-defined algorithms to return the results expected by the user. It can usually be applied to fields such as text retrieval, machine translation, and information question answering. Users usually define the algorithm by establishing an algorithm model, and the established algorithm model needs to be trained through a large number of labeled original language materials; labeling the original language materials refers to processing the original corpus, and combining various representations of language features The additional codes are marked on the corresponding language component...

AI Technical Summary

Problems solved by technology

[0003] However, the inventors found that there are at least the following problems in the related technologies: in the traditional corpus labeling method, a single corpus is generally marked by a single labeler, and the cognitive level and operating habits of the labeler greatly affect the quality of labeling the corpus. As a result, the labeling results of the corpus are single, and it is difficult to judge the accuracy of the labeling results

Images