Text sentence segmentation method and device, electronic device and storage medium

Abstract

Embodiments of the invention provide a text sentence segmentation method and device, an electronic device and a storage medium. The method comprises the steps of determining a character feature vectorof each character in a text; inputting the character feature vector of each character into a sentence segmentation model to obtain a sentence segmentation probability of each character output by thesentence segmentation model; wherein the sentence segmentation model is obtained by training based on a sample word feature vector and a sentence segmentation identifier of a sample word in the sampletext; determining a plurality of candidate sentence segmentation results based on the sentence segmentation probability of each character; and determining a sentence segmentation result based on a preset word number threshold and the plurality of candidate sentence segmentation results. According to the method and device, the electronic device and the storage medium provided by the embodiment ofthe invention, the sentence segmentation result that the length of each clause is smaller than or equal to the preset word number threshold value is obtained while the local semantics are not cut off,so that efficient and accurate text sentence segmentation is realized, and the loss of labor cost and time cost is avoided.

Description

technical field [0001] The present invention relates to the technical field of natural language processing, in particular to a text segmentation method, device, electronic equipment and storage medium. Background technique [0002] Subtitles refer to the explanatory text displayed on the playback interface when the audio and video are playing, which can help viewers understand the audio and video content. [0003] At present, sentence segmentation of subtitle texts is mostly done manually, which consumes a lot of human resources and time. Although the rapid development of natural language processing technology has made the text segmentation technology more and more mature, there are usually specific text segmentation requirements when segmenting subtitle texts, and the general text segmentation technology cannot meet the segmentation requirements of subtitle texts. Contents of the invention [0004] Embodiments of the present invention provide a text segmentation method, ...

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a text segmentation method, device, electronic equipment, and storage medium to solve the problem that the existing subtitle text segmentation relies on manual completion, which takes time and effort

Images