Text element extraction method

Abstract

The invention discloses a text element extraction method, which comprises the following steps of: segmenting a text, and converting the segmented text into a character sequence; assigning a plurality of representation modes to the characters, and combining to obtain a vector set; dividing the vector set into a plurality of subsets and carrying out element extraction model training for a plurality of times to obtain a final model; and performing text element extraction by utilizing the final model according to the matching rule. Each character has multiple representation modes, so that the data volume of the vector set is huge, and meanwhile, data resources are fully utilized, the data volume requirement is reduced and the efficiency is improved through the split sub-set form, mutual verification and optimization. The method has the substantive effects that the data acquisition problem of the newly-added elements in the irregular mode can be solved, the manual labeling cost is reduced, the newly-added elements with the regular mode can be directly added on an existing model to serve as a new optimization model, the service implementation time is shortened, and the recognition effect is enhanced.

Description

technical field [0001] The invention relates to the field of data processing, in particular to a method for extracting text elements. Background technique [0002] Named entity recognition aims to extract named entities from text and classify them into different categories. Common identification categories include personal names, place names, and organization names. For example, the invention of authorized announcement number CN102750390B discloses a method for automatically extracting news webpage elements, which can automatically identify news elements. Named entities are one of the most important fields in the field of natural language processing, and they are also the prerequisites for many downstream application tasks or the pre-tasks to improve their accuracy, such as entity linking, both of which are very important for building large-scale knowledge graphs. [0003] Machine learning refers to the process of using certain algorithms to guide the computer to use known ...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problems of low processing efficiency and large data volume demand in the element extraction method in the prior art, the present invention provides a text element extraction method, which uses limited text data to split and combine the generated character sequences and vector sets, etc. Reuse of methods, improve data utilization efficiency, and reduce data volume requirements