Trade name recognition method based on full-text attention mechanism

A technology of commodity name and identification method, which is applied in neural learning methods, special data processing applications, instruments, etc., to achieve the effect of improving accuracy
CN109062893AActive Publication Date: 2018-12-21SOUTH CHINA UNIV OF TECH
5 Cites 28 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Applications(China)
Current Assignee / Owner
SOUTH CHINA UNIV OF TECH
Publication Date
2018-12-21

Smart Images

  • Figure 1
    Figure 1
Patent Text Reader

Abstract

The invention discloses a commodity name recognition method based on a full-text attention mechanism. The method comprises the following steps: establishing a depth neural network model with a full-text attention mechanism; adding a full-text attention layer into the depth neural network model; in the full-text attention layer, the full-text information being represented by a method of embedding aword into a vector; the full-text attention layer being represented by a full-text information embedding vector. Then, for each word to be annotated, the similarity between them and the full-text attention-level word vector is calculated, so that the word can obtain the 'attention weight' of the full-text, that is, the word can focus on the context information of the document, and the additionalfeatures of each word can be calculated by using the sum of the attention-level weight, so as to identify the trade name. The method of the invention can effectively extract commodity names from irregular texts, solves the problem of inconsistent identification of the same commodity in context, and improves the accuracy of identification.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical field of natural language processing, in particular to a product name recognition method based on a full-text attention mechanism. Background technique

[0002] Named entity recognition is a subtask of information extraction that aims to identify proper nouns in text and classify them. Traditional named entity tasks include: person names, place names, organization names, etc. This is a very important basic task in the field of natural language processing, such as: search engines, question answering systems, recommendation systems, translation systems, etc. In recent years, with the continuous development of e-commerce, the task of named entity recognition in the field of e-commerce has also begun to attract people's attention. People urgently need to extract product names from massive and irregular texts and use them In intelligent customer service, advertising recommendation and other fields.

[0003] For this...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More