A Boundary Composition Named Entity Recognition Method Based on Neural Network

Abstract

The invention discloses a neural network-based boundary combination named entity recognition method, comprising the following steps: step 1: extracting entity boundary information based on a neural network model, and constructing a boundary recognition model; Combine to obtain candidate entity sets; Step 3: Build a neural network classifier to screen candidate entity sets. The method disclosed in the present invention adopts the boundary combination strategy, introduces neural network technology, fully utilizes the characteristics of neural network layered automatic extraction of high-dimensional abstract features, divides entity recognition into three steps of boundary recognition, boundary combination and candidate entity recognition, and makes up for It overcomes the shortcomings of the traditional sequence model, and to a certain extent, avoids the feature sparsity problem caused by the traditional machine learning method, thereby improving the performance of nested named entity recognition and achieving good results.

Description

technical field [0001] The invention relates to a named entity recognition method, in particular to a neural network-based boundary combination named entity recognition method, belonging to the technical fields of natural language processing and machine learning. Background technique [0002] With the popularity of computers and the rapid development of the Internet, a large amount of information appears in front of people in the form of electronic documents. In order to cope with the severe challenges brought by the information explosion, there is an urgent need for professional automated tools to extract truly valuable information from massive amounts of data, and information extraction has emerged as the times require. Named entities refer to the proper nouns in the text that represent the names of people, places, and organizations. As an important semantic knowledge carrier in the text, named entity recognition plays an important role in information extraction. After it ...

AI Technical Summary

Problems solved by technology

Sequence models set tags through each character at the sentence level to obtain the most probable labeling path, but cannot effectively identify internal nested entities; grammatical analysis is identified by using a grammatical analysis tree, but often depends on the effect of grammatical analysis; based on embedding The nested model can better deal with the nesting problem of named entity recognition

However, these methods have four shortcomings: first, they are all in the sentence expansion task, and there is a problem of sparse features; second, in the sequence model, changing the annotations of internal (or external) entities will not be conducive to feature weighting; third, Treating different classes separately will not be able to effectively use tag information; finally, entity recognition is affected to a certain extent by cascading errors caused by word segmentation or grammatical parsing

Images