This article presents a relations taxonomy constructed on the basis of a conceptual framework, as well as methods of identifying semantic relations that are used for automatically constructing ontologies of subject areas. A predicate lexicon that corresponds to relation types is established through automatically processing a collection of scientific and technical document texts. Patterns of linguistic constructions of semantic relations are constructed in a way that considers morphological features of words and morphemes. Statistics of types of relations and the use of patterns depending on the document type are studied.

中文翻译：

---

科技主题领域本体中语义关系的系统化与识别

本文介绍了基于概念框架构建的关系分类法，以及识别用于自动构建主题领域本体的语义关系的方法。通过自动处理科学和技术文档文本的集合来建立与关系类型相对应的谓词词典。语义关系的语言构造模式的构建考虑了单词和词素的形态特征。研究了关系类型的统计数据以及根据文档类型使用的模式。