融合分词信息的中文命名实体识别模型
Chinese NER model of fusing word segmentation information
摘要
中文命名实体识别在中文自然语言处理技术中有相当重要的作用。为了解决字符与词汇信息难统一和现有方法较为复杂的问题，本文提出了一种融合分词信息的模型。首先设计了通用的多任务学习方式，使模型同时进行命名实体识别任务与分词任务的学习。然后设计了特征融合模块，将命名实体识别特征与分词特征相融合。该模型在3个中文数据据集上进行了实验，实验结果表明了本文所提方法能有效提高实体识别的F1值。
Abstract
Chinese named entity recognition plays an important role in Chinese natural language processing.In order to solve the problem that the information of characters and words is difficult to unify and the existing methods are more complex, this paper proposes a model of fusing word segmentation information.Firstly, a general multi-task learning method is designed to make the model learn named entity recognition task and word segmentation task at the same time. Then, the feature fusion module is designed to fuse the named entity recognition feature with the word segmentation feature. Experiments on three Chinese named entity recognition datasets show that the proposed method can effectively improve the F1 score of entity recognition.关键词
人工智能/中文命名实体识别/中文分词/多任务学习?Key words
ArtificialIntelligence/Chinese Named Entity Recognition/Chinese Word Segmentation/Multi-task Learning引用本文复制引用
肖波,陈柯宏.融合分词信息的中文命名实体识别模型[EB/OL].(2021-03-10)[2026-04-04].http://www.paper.edu.cn/releasepaper/content/202103-105.学科分类
汉语
评论