Text processing method, device and equipment

Abstract

The embodiment of the invention discloses a text processing method, device and equipment. The method comprises the steps of obtaining a to-be-recognized target text; performing grammatical relationship analysis on the target text to obtain a target grammatical relationship sequence corresponding to the target text; matching the target grammar relationship sequence with each reference sequence modein a reference sequence mode set, and selecting a matching sequence mode matched with the target grammar relationship sequence; and obtaining a target grammatical relationship identifier behind the mth reference grammatical relationship identifier in the target grammatical relationship sequence, and determining a character corresponding to the target grammatical relationship identifier in the target text as a name character. According to the embodiment of the invention, the name entity can be better recognized from the text, and the name recognition accuracy is improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a text processing method, a text processing device and a text processing device. Background technique [0002] Many application scenarios (for example, portrait construction scenarios, map construction scenarios, event mining scenarios, public opinion analysis scenarios, etc.) involve name recognition. The so-called name recognition can refer to the process of identifying the name entity in the text (or field) to be recognized. [0003] In the prior art, the methods for carrying out name recognition mainly include the following: 1. The name recognition method based on sequence labeling, which regards the name recognition problem as a simple sequence labeling problem; this simple labeling method , making the accuracy of name recognition lower. ②Name recognition method based on recurrent neural network, this recognition method is based on the premise of complex networ...

AI Technical Summary

Problems solved by technology

[0003] In the prior art, the methods for carrying out name recognition mainly include the following: 1. The name recognition method based on sequence labeling, which regards the name recognition problem as a simple sequence labeling problem; this simple labeling method , making the accuracy of name recognition lower

②Name recognition method based on recurrent neural network, this recognition method is based on the premise of complex network training, the efficiency is low, and it depends on manually specifying context rules, which has poor flexibility, low recall rate and transplantation Poor performance and other issues

③The name recognition method based on the statistical model. This recognition method is based on the ngram (that is, an algorithm based on the statistical language model) model for name recognition, which cannot provide sufficient reference features, and may easily cause misjudgment of the name entity. less accurate

To sum up, it can be seen that the existing methods for name recognition have the problem of low accuracy. Therefore, how to accurately identify the name entity from the text to be recognized has become an important topic of research.

Images