A Chinese grammatical error detection method based on word vectors added with text information

Abstract

The invention discloses a Chinese grammatical error detection method and device adding word vectors of text information, belonging to the field of information processing. The characteristics of this method include: first vectorize the words of the input text to form a text matrix; then use the cyclic neural network to form text information related to each word vector; reconstruct the text matrix; use the cyclic neural network to extract context information; Compute error scores for each word to the neural network; use the error scores to infer error locations. The present invention improves the detection effect of Chinese grammar by combining text-based word vectors, and has great use value.

Description

technical field [0001] The invention relates to the field of information processing, in particular to a neural network-based Chinese grammatical error detection method. Background technique [0002] Due to the rapid development of China, more and more foreigners start to learn Chinese, so the task of Chinese grammatical error detection has attracted more and more attention. The purpose of the Chinese grammatical error detection task is to judge whether there are grammatical errors in the text written by non-Chinese native speakers, and to give an error message. [0003] Most of the current grammatical error detection models use sequence labeling. The model marks the wrong words in the text through calculation and gives an error message. Commonly used statistical learning methods for Chinese grammatical error diagnosis include n-gram, and machine learning methods include recurrent neural network methods. However, these networks require more artificial features to achieve be...

AI Technical Summary

Problems solved by technology

At present, most of the existing detection methods do not make good use of the information expressed in Chinese vocabulary, ignoring the phenomenon of polysemy in Chinese

Images