| 王昊,严明,苏新宁.基于机器学习的中文书目自动分类研究[J].中国图书馆学报,2010,36(6):0 |
| 基于机器学习的中文书目自动分类研究 |
| Research on Automatic Classification for Chinese Bibliography Based on Machine Learning |
| |
| DOI： |
| 中文关键词: 机器学习，书目自动分类，特征加权，中图法，浅层次分类模型 |
| 英文关键词: Machine learning,Automatic bibliography classification,Feature weighted,Chinese Library Classification,Shallow classification model |
| 基金项目:本文系国家社科基金项目“面向语义网本体的知识管理研究”（编号：09CTQ010）的研究成果之一。 |
|
| 摘要点击次数: 5270 |
| 全文下载次数: 4429 |
| 中文摘要: |
| 面对与日俱增的图书出版量，图书馆编目人员的手工书目分类显得力不从心，如何实现由计算机自动完成图书分类成为数字图书馆建设中亟待解决的关键问题之一。本文尝试将BP神经网络和支持向量机等机器学习算法引入到书目分类中，建立了面向中图法的基于机器学习的书目层次分类系统模型，提出了采用特征加权方式描述书目和浅层次分类体系构建的设计思路，并通过大规模实验验证了该模型的可行性和合理性，基本上解决了没有主题标注情况下书目的自动分类问题。图9。表5。参考文献14。 |
| 英文摘要: |
| Books classification by computer has become one of the most critical issues which should been solved immediately in digital library construction because of increasing volume of book publishing. This paper tries to induct the BP nerve net and Support Vector Machine algorithms to bibliography classification, and establishs bibliography hierarchy classification system model based on machine learning faced to the Chinese Library Classification, then proposes the design ideas of describing bibliographies using feature weighted mode and constructing shallow classification system. It verifies the feasibility and rationality of the model by large scale experiment, and basically solves the case of the bibliography automatic classification without subject labeling, which lays a theoretical foundation for constructing the practical bibliography automatic classification system, and provides factual basis for the wide range application of machine learning methods for the construction of digital libraries. 9 figs. 5 tabs. 14 refs. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
