arXivDaily arXiv每日学术速递 周一至周五更新
arXiv cs/0610058cs.IR
收藏

上下文相关的电子文档语料库访问

Context-sensitive access to e-document corpus

A. V. Smirnov, T. V. Levashova, M. P. Pashkin, N. G. Shilov, A. A. Krizhanovsky, A. M. Kashevnik, A. S. Komarova

摘要:

上下文相关的电子文档访问方法将上下文作为一个问题模型,从应用领域中提取知识,并以应用本体的形式表示。需要高效地访问文本形式的信息。Wiki资源作为一种现代文本格式,以半形式化的结构提供了大量的文本。在该方法的第一阶段,对代表宏观情况的本体对文档进行索引。索引方法使用主题树作为文档和应用本体之间的中间层。在第二阶段,确定与当前情况有关的文件(抽象和操作背景),并按相关程度进行排序。抽象上下文是一个面向问题的基于本体的模型。操作上下文是抽象上下文与信息源提供的数据的实例化。描述了该方法的以下部分:(i)用于测量电子文档与本体的相似性的度量,(ii)存储针对本体的电子文档的索引的结果的文档索引;(iii)用于基于语义相似性度量来识别相关电子文档的方法。维基百科(维基资源)是作为一个语料库的电子文档的方法评估的案例研究。文本分类、元数据的存在以及与不同主题相关的大量文章的存在是语料库的特征。

英文摘要:

The methodology of context-sensitive access to e-documents considers context as a problem model based on the knowledge extracted from the application domain, and presented in the form of application ontology. Efficient access to an information in the text form is needed. Wiki resources as a modern text format provides huge number of text in a semi formalized structure. At the first stage of the methodology, documents are indexed against the ontology representing macro-situation. The indexing method uses a topic tree as a middle layer between documents and the application ontology. At the second stage documents relevant to the current situation (the abstract and operational contexts) are identified and sorted by degree of relevance. Abstract context is a problem-oriented ontology-based model. Operational context is an instantiation of the abstract context with data provided by the information sources. The following parts of the methodology are described: (i) metrics for measuring similarity of e-documents to ontology, (ii) a document index storing results of indexing of e-documents against the ontology; (iii) a method for identification of relevant e-documents based on semantic similarity measures. Wikipedia (wiki resource) is used as a corpus of e-documents for approach evaluation in a case study. Text categorization, the presence of metadata, and an existence of a lot of articles related to different topics characterize the corpus.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑