国家科学图书馆青年人才领域前沿项目结题报告.DOC

PAGE 国家科学图书馆青年人才领域前沿项目结题报告 基于本体的科学知识图谱 分析方法研究 撰写人：王建芳、吴清强、张超星、安新颖 撰写时间：2011年6月10日 摘 要 由于目前基于共词关系绘制的科学图谱可读性不强的问题，本研究尝试利用本体的语义关系来优化共词科学图谱的关系展示，通过理论分析、实验案例来揭示其可行性和有效性。 本研究借鉴相关领域的研究成果，提取本体关系中的上下位关系及关联关系计算词与词之间的语义关系，并根据两种类型的语义关系构建了其与共词关系融合的算法。通过“农业污染”领域的案例实验，分析了共词关系与本体中体现的上下位关系、关联关系融合的效果，结果显示，上下位关系对共词关系图谱的优化效果有效，而关联关系可以在一定程度上丰富语义关系，并优化聚类结果。 关键词：科学知识图谱；本体；关系融合 Abstract As a method of science-mapping，co-word based mapping always have a question of readability. This study attempted to use semantic relations to optimize the co-word based sicence mapping. The semantic relations are defined by ontology. we use theoretical analysis and test case to reveal the feasibility and effectiveness . In this study, drawing on research in related fields, we extracted hyponymy relations and “related” relations between words from the selected ontology. Then based on the two types of semantic relations, built the integration algorithm between semantic relations and co-word relations. Through the agricultural pollution field case experiments, the results show that hyponymy relations contributes few to the results of co-word scicene mapping, while “related” relations is better, and after relaiton integration, we can see richer semantic relations to some extent, and optimized the clustering results. Keywords: science mapping; ontology; relation integration 目 录 TOC \o 1-3 \h \z \u HYPERLINK \l _Toc298997938 一、研究背景与现状 PAGEREF _Toc298997938 \h 1 HYPERLINK \l _Toc298997939 1.1研究背景 PAGEREF _Toc298997939 \h 1 HYPERLINK \l _Toc298997940 1.2国内外研究进展 PAGEREF _Toc298997940 \h 2 HYPERLINK \l _Toc298997941 1.2.1科学知识图谱及相关研究进展 PAGEREF _Toc298997941 \h 2 HYPERLINK \l _Toc298997942 1.2.2 基于本体的情报分析研究进展 PAGEREF _Toc298997942 \h 4 HYPERLINK \l _Toc298997943 二、科学图谱语义解释性问题及引入本体的可行性 PAGEREF _Toc298997943 \h 6 HYPERLINK \l _Toc298997944 2.1科学图谱语义解释性问题的分析 PAGEREF _Toc298997944 \h 6 HYPERLINK \l _Toc298997945 2.2基于共词的科学图谱中引入本体关系的可行性分析 PAGEREF _Toc298997945 \h 7 HYPERLINK \l _To