基于BERT的《中图法》文本分类系统及其影响因素分析  

A Case Study of the BERT Model Based on Chinese Library Classification and Influence Factors

作　　者：姜鹏 Jiang Peng

机构地区：[1]上海图书馆,上海200031

出　　处：《图书馆研究与工作》2022年第5期43-48,共6页Library Science Research & Work

摘　　要：文章以BERT模型为例,在分析了BERT模型分类结果的基础上,总结了基于《中国图书馆分类法》的文本分类经验。从数据、系统、分类法等方面详细分析了目前影响分类系统的因素,并通过训练数据集清洗、引入知识图谱等措施来提高分类准确率。Taking the BERT model as an example,this paper tries to analyze the classification results based on the model and conclude the text classification based on the Chinese Library Classification(CLC).The paper also analyzes influence factors of classification from the perspectives of data,systems and classification methods.What’s more,the author of this paper aims to make the classification results more accurate through data cleaning and the knowledge map.

关 键 词：文本分类 《中国图书馆分类法》 BERT 

分 类 号：G254.1[文化科学—图书馆学] TP18[自动化与计算机技术—控制理论与控制工程]