首页 | 本学科首页   官方微博 | 高级检索  
检索        

基于Labeled LDA主题模型的医学文献自动分类法
引用本文:宫小翠,安新颖,单连慧.基于Labeled LDA主题模型的医学文献自动分类法[J].中华医学图书情报杂志,2018,27(10):53-58.
作者姓名:宫小翠  安新颖  单连慧
作者单位:中国医学科学院医学信息研究所,北京 100020,中国医学科学院医学信息研究所,北京 100020,中国医学科学院医学信息研究所,北京 100020
基金项目:中国医学科学院中央级公益性科研院所基本科研业务费专项“科技创新环境下医学科研机构科技成果转化能力评价研究”(2017PT63004);国家重点研发计划子课题“精准医学本体语料库构建与扩展”(2016YFC0901902-2)
摘    要:提出了一种基于Labeled LDA主题模型的医学文献自动分类法。以10个医学领域的研究文献为案例,通过语料库的设置及参数设置调整模型为最佳,与SVM方法进行对比实验。结果显示,无论是准确率还是召回率,基于Labeled LDA主题模型的自动分类法均比SVM法高出7.00%左右,表明基于Labeled LDA主题模型的医学文献自动分类法具有较好的医学领域文本分类效果。

关 键 词:Labeled  LDA  主题模型  自动分类  SVM
收稿时间:2018/9/30 0:00:00

Labeled LDA topic model-based automatic classification of medical literature
AN Xin-ying,AN Xin-ying and SHAN Lian-hui.Labeled LDA topic model-based automatic classification of medical literature[J].Chinese Journal of Medical Library and Information Science,2018,27(10):53-58.
Authors:AN Xin-ying  AN Xin-ying and SHAN Lian-hui
Institution:Institute of Medical Information, Chinese Academy of Medical Sciences, Beijing 100020, China,Institute of Medical Information, Chinese Academy of Medical Sciences, Beijing 100020, China and Institute of Medical Information, Chinese Academy of Medical Sciences, Beijing 100020, China
Abstract:A Labeled LDA topic model-based automatic classification of medical literature was proposed. The Labeled LDA topic model was adjusted to its best performance by establishing its language repository and setting its parameters with the literature in 10 medical fields as an example and compared with the SVM-based classification of medical literature, which showed that the precision rate and recall rate of Labeled LDA topic model-based automatic classification were about 7.00% higher than those of SVM-based classification, indicating that the efficacy of Labeled LDA topic model-based automatic classification is higher than that of SVM-based classification for the classification of medical literature.
Keywords:Labeled LDA  Topic model  Automatic classification  SVM
点击此处可从《中华医学图书情报杂志》浏览原始摘要信息
点击此处可从《中华医学图书情报杂志》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号