Text recognition method and device based on Chinese word vector model and storage medium

Abstract

The invention provides a text recognition method and device based on a Chinese word vector model and a storage medium. The method comprises the following steps: constructing a first Chinese word vector model and a second Chinese word vector model based on a target word, a context window word and a target word tone n-gram; training the first Chinese word vector model and the second Chinese word vector model based on a corpus to obtain the trained first Chinese word vector model and second Chinese word vector model; and obtaining a semantic vector of an input text by using the trained first Chinese word vector model and outputting a result. According to the invention, the trained second Chinese word vector model is used to verify the identification result; the recognition accuracy is improved, the tone element is used when the word vector is constructed, the information contained in the Chinese word vector is enriched, the influence of feature noise on the word vector is avoided, the semantic and intonation features of the target word are reserved to the maximum extent, a negative sample is introduced, and a loss function is constructed.

Description

technical field [0001] The invention relates to the technical field of text data processing, in particular to a text recognition method, device and storage medium based on a Chinese word vector model. Background technique [0002] As one of the basic works of natural language processing, text representation has been attracting much attention. Through text representation technology, unstructured text can be converted into structured data that computers can process, thereby helping computers to carry out subsequent work. Traditional text representation is realized based on the idea of ​​probability and statistics, typical representatives are bag-of-words model and topic model; the former is difficult to achieve good results in practical tasks due to the problems of ignoring semantic information and data sparsity, while the latter Because of the limitation of prior knowledge, it is also difficult to be generalized in downstream tasks. [0003] With the development of deep lea...

AI Technical Summary

Problems solved by technology

Considering that the semantics contained in Chinese characters and the semantics contained in words are redundant, the joint training of words and words cannot guarantee the complementarity of semantic information; and Pinyin, as a tool to assist Chinese character learning, itself is a cluster gathered by rules, and the information it contains It does not have a direct meaning for understanding text semantics. Blindly combining multiple character items will often introduce more noise as a result, which will bring difficulties to the learning of subsequent models.

[0006] Therefore, the Chinese word vector model in the prior art has a lot of noise, complicated and time-consuming training, and the recognition of text is not necessarily effective, the convergence speed of the model is slow during training, etc., and a new Chinese word vector model is urgently needed Addresses one or more of the above technical deficiencies

Images