面向维基百科的领域知识演化关系抽取
发布时间：2018-09-04 18:20
【摘要】：互联网下同一领域中不同知识概念间存在多种关系,其中演化关系对于用户学习和理解领域知识,梳理领域知识的前序和后续逻辑关系具有重要意义,然而网络数据的多样和无序使用户难以准确有序地获取领域知识关系.针对该问题,提出一种面向中文维基百科领域知识的演化关系抽取方法,利用语法分析特征,挖掘演化关系模式,构建演化关系推理模型,采用基于句子层面的关系抽取算法识别领域知识演化关系,最后在真实的维基百科数据集上对该文方法进行了性能评测.实验表明,该方法具有较高的关系抽取准确率和召回率,能有效地抽取出维基百科中领域知识的演化关系.同时,基于实验抽取结果构建知识图谱,能有效挖掘领域学科下知识集合的演化体系,识别重难点知识,对学科建设以及相关课程教学具有一定的指导意义.
[Abstract]:There are many relationships between different concepts of knowledge in the same domain under the Internet. Among them, evolutionary relationships are of great significance for users to learn and understand domain knowledge, to sort out the preorder and subsequent logical relationships of domain knowledge. However, the diversity and disorder of network data make it difficult for users to acquire domain knowledge relationship accurately and orderly. In order to solve this problem, an evolutionary relationship extraction method for Chinese Wikipedia domain knowledge is proposed, which utilizes the features of grammar analysis, mining evolutionary relational patterns, and constructing evolutionary relational reasoning model. The relationship extraction algorithm based on sentence level is used to identify the evolutionary relationship of domain knowledge. Finally, the performance of this method is evaluated on the real Wikipedia dataset. Experiments show that the proposed method has a high accuracy and recall rate of relation extraction, and it can effectively extract the evolutionary relationship of domain knowledge in Wikipedia. At the same time, constructing the knowledge map based on the experimental results can effectively mine the evolution system of knowledge set under the domain discipline, identify the important and difficult knowledge, and have certain guiding significance for the subject construction and related course teaching.
【作者单位】： 西南科技大学计算机科学与技术学院;西南科技大学教育信息化推进办公室;中国科学技术大学计算机科学与技术学院;
【基金】：四川省教育厅资助项目(14ZB0113) 西南科技大学博士基金(12zx7116) 赛尔网络下一代互联网技术创新项目(NGII20150510)资助~~
【分类号】：TP391.1
本文编号：2222971
[Abstract]:There are many relationships between different concepts of knowledge in the same domain under the Internet. Among them, evolutionary relationships are of great significance for users to learn and understand domain knowledge, to sort out the preorder and subsequent logical relationships of domain knowledge. However, the diversity and disorder of network data make it difficult for users to acquire domain knowledge relationship accurately and orderly. In order to solve this problem, an evolutionary relationship extraction method for Chinese Wikipedia domain knowledge is proposed, which utilizes the features of grammar analysis, mining evolutionary relational patterns, and constructing evolutionary relational reasoning model. The relationship extraction algorithm based on sentence level is used to identify the evolutionary relationship of domain knowledge. Finally, the performance of this method is evaluated on the real Wikipedia dataset. Experiments show that the proposed method has a high accuracy and recall rate of relation extraction, and it can effectively extract the evolutionary relationship of domain knowledge in Wikipedia. At the same time, constructing the knowledge map based on the experimental results can effectively mine the evolution system of knowledge set under the domain discipline, identify the important and difficult knowledge, and have certain guiding significance for the subject construction and related course teaching.
【作者单位】： 西南科技大学计算机科学与技术学院;西南科技大学教育信息化推进办公室;中国科学技术大学计算机科学与技术学院;
【基金】：四川省教育厅资助项目(14ZB0113) 西南科技大学博士基金(12zx7116) 赛尔网络下一代互联网技术创新项目(NGII20150510)资助~~
【分类号】：TP391.1
【相似文献】
相关期刊论文 前1条
1 曾蕴波;王明微;范海涛;;面向MBSE的产品阶段模型的要素演化关系发现方法研究[J];制造业自动化;2013年05期
，本文编号：2222971
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2222971.html
