Chinese named entity recognition method and system based on pen painting convolution and word vectors

Abstract

The invention provides a Chinese named entity recognition method and system based on pen and picture convolution and word vectors, and relates to the technical field of named entity recognition, and the method comprises the steps: obtaining a stroke sequence corresponding to each Chinese character in a text and a character feature vector of each Chinese character; inputting the stroke sequence into a pen-painting convolutional neural network to obtain a stroke feature vector; setting a sliding window according to the maximum length of an entity in the text, and obtaining a word vector of each character in the sliding window through a self-attention mechanism; splicing the stroke feature vector, the word vector and the character feature vector of each Chinese character in the text, and inputting into the BiLSTM network to obtain the score of each Chinese character corresponding to each entity tag; and determining an optimal entity label for each Chinese character in the text by adopting a CRF model. According to the method, the influence of the stroke sequence of the Chinese character on the Chinese character is considered, and the named entity recognition is performed after the stroke feature vector, the word feature vector and the character feature vector of the Chinese character are combined, so that the named entity recognition effect is improved.

Description

technical field [0001] The present invention relates to the technical field of named entity recognition, in particular to a Chinese named entity recognition method and system based on stroke convolution and word vectors. Background technique [0002] With the rapid development of Internet technology and the continuous growth of unstructured data, we are in an era of massive and unstructured data. How to efficiently manage data and extract effective information from unstructured data has become an urgent problem to be solved. [0003] The purpose of Named Entity Recognition (NER) is to identify defined named entities from unstructured text, such as person names, place names, organization names, etc., which is the basic core task of information retrieval and information extraction. Chinese NER is a division of NER in the Chinese field. Due to the characteristics of Chinese characters, there are still many problems in Chinese NER. The main difficulties of Chinese NER are as f...

AI Technical Summary

Problems solved by technology

The word-based tagging method first uses the word segmentation tool to segment the text, and then performs entity recognition. The word boundary of this type of method is also an entity boundary. If an error occurs in the word segmentation stage, the subsequent NER model will not be able to correctly identify the entity.

The word-based sequence tagging method usually has insufficient semantics, so people mainly consider how to make better use of word information. Some applications introduce external lexical information based on the word-based sequence tagging method and integrate it into the word vector at the input layer. In the representation, this makes the model itself changed, and the introduction of external word vectors also makes the model training efficiency low, and finally the accuracy of named entity recognition is reduced; The ElMo model established by the stroke sequence has defects in the effectiveness and accuracy of named entity recognition

Images