基于深度学习的中文命名实体识别技术研究
Research on Chinese Named Entity Recognition Technology Based on Deep Learning
武文静 1岳杰 1王佳丽 1刘枫1
作者信息
- 1. 河北建筑工程学院,河北 张家口 075000
- 折叠
摘要
命名实体识别(NER)是 NLP领域的一项基础底层任务.针对当前传统的基于规则和统计方法存在特征提取的精准度和模型的可扩展性上不足的问题,中文命名实体识别技术在利用神经网络学习模型时得到了极大地改善.除了通过 Bert预训练模型和相关的公开数据集对文本数据特征提取、识别实体之外还融合了人工标注的地名和组织机构实体的额外数据集来增强模型的词义理解准确度.实验结果表明,模型的实体识别能力有所提高.
Abstract
Named Entity Recognition(NER)is a fundamental underlying task in the field of NLP.In re-sponse to the shortcomings of traditional rule-based and statistical methods in feature extraction accuracy and model scalability,Chinese named entity recognition technology has greatly improved when using neural networks to learn models.In addition to using Bert pre trained models and rele-vant public datasets for text data feature extraction and entity recognition,this article also in-tegrates additional datasets of manually annotated place names and organizational entities to en-hance the accuracy of the mode'ls semantic understanding.The experimental results show that the entity recognition ability of the model has been improved.
关键词
自然语言处理/中文命名实体识别/深度学习/中文分词Key words
natural language processing/Chinese Named Entity Recognition/Deep Learning/Chinese word segmentation引用本文复制引用
出版年
2024
NETL
NSTL
万方数据