Identification method and device based on text enhancement, electronic equipment and storage medium

Abstract

The invention relates to the field of artificial intelligence, and provides a text enhancement-based recognition method, which comprises the following steps of: firstly, converting a pre-constructed program statement into an English statement, and performing first enhancement processing on the English statement according to a preset English standardization processing rule to form an English sentence set; then translating the English sentence set into a Chinese sentence set through a preset translation plug-in, performing second enhancement processing on the Chinese sentence set according to a preset Chinese processing rule to form text samples, then splicing the text samples to form sample data, and inputting the sample data into a preset to-be-converged model to obtain a text recognition model; according to the method, the to-be-recognized text is input into the text recognition model for text recognition processing to obtain the named entity recognition information or the text recognition information about the to-be-recognized text, so that the effective sample set can be quickly and specifically generated, the richness of the text sample is improved, and the accuracy of entity labeling is improved.

Description

technical field [0001] The present invention relates to the field of artificial intelligence, relates to a text enhancement method, in particular to a text enhancement-based recognition method, device, electronic equipment and computer-readable storage medium. Background technique [0002] Named entity recognition is one of the most basic tasks of natural language processing tasks, and one of the most important basic tasks of tasks such as reading comprehension, dialogue systems, and machine translation. Training a model capable of named entity recognition requires a large amount of sample data. Therefore, data enhancement is first required to generate a large number of text samples to achieve model training. [0003] In computer vision, data augmentation is a very common and simple method of enriching samples, such as rotating, zooming in, mirroring, converting to grayscale, etc., without changing its semantics. But in natural language processing, the data augmentation ope...

AI Technical Summary

Problems solved by technology

[0006] The invention provides a text enhancement-based recognition method that can quickly and targetedly generate an effective sample set to increase the richness of text samples and further improve the accuracy of entity labeling, so as to solve the problem of data enhancement in natural language processing. The operation must be very cautious, otherwise the semantics of the text will be easily changed, resulting in poor model results; in the existing field of natural language processing, how to solve the dilemma of small samples, proposed unconditional enhancement, conditional enhancement, semi-supervised learning , UDA and other solutions, but when it is implemented, the professionalism of the existing data sets is not strong enough, and the generated samples still do not meet the problem of named entity recognition.

Images