Brand name identification method, computer device and computer readable storage medium

Abstract

The invention provides a brand name identification method, a computer device and a computer readable storage medium. The brand name identification method comprises the following steps of setting a plurality of labels; obtaining a training text, obtaining a preset brand name from the training text according to the set labeling label, and labeling the preset brand name; obtaining a representation vector of each character in the training text, inputting a character string composed of a plurality of characters into a BERT model, obtaining the representation vector of each character, calculating the probability of each label of each character, using the label with the maximum probability as a target label of the character, and training a brand recognition model; And obtaining a to-be-identifiedtext, inputting the to-be-identified text into the brand identification model, obtaining a representation vector of each character, and calculating a predicted brand name according to the label of each character of the to-be-identified text. The invention also provides a computer device for realizing the brand name identification method and a computer readable storage medium. According to the present invention, the brand name recognition accuracy can be improved, and the manual participation degree is reduced.

Description

technical field [0001] The invention relates to the field of computer character recognition, in particular to a brand name recognition method, a computer device for realizing the method, and a computer-readable storage medium. Background technique [0002] With the popularization of smart electronic devices, people increasingly use smart electronic devices to communicate, and various social software and information sharing websites have become platforms for people to share their life, work, various experiences and emotions. With the development of people's application of various social software and information sharing websites, a huge social network has been formed, and people can keep abreast of the information of their friends and celebrities they follow through the social network. [0003] With the continuous expansion of the scale of social networks, the analysis of text information in social networks has become the main data source for analyzing user needs and inferring...

AI Technical Summary

Problems solved by technology

However, the disadvantages of these three methods are that they require the participation of expert knowledge, many features are manually defined, and they cannot be used universally across domains.

[0005] Moreover, due to the complexity and diversity of brand names, and the lack of domain-related training corpus, it often takes a lot of labor costs to collect and label the corpus, and the colloquial and random characteristics of social network texts also give rise to problems. Accurate identification of brand names adds a lot of difficulty

Images