Retrieval in text collections with historic spelling using linguistic and spelling variants

Retrieval in text collections with historic spelling using linguistic and spelling variants
复制标题

使用语言和拼写变体检索具有历史拼写的文本集合

DOI:
--
复制
发表时间:
2007
期刊:
ACM/IEEE Joint Conference on Digital Libraries
影响因子:
--
通讯作者:
N. Fuhr
N. Fuhr
中科院分区:
--
文献类型:
--
作者:
Andrea Ernst;N. Fuhr

文献摘要

被引文献

相似文献

我们提出了一种新的方法检索文本与非标准的拼写,这是很重要的历史文本,例如在英语或德语。在本文中,我们描述了我们的系统的整体架构,其次是它的评价。给定一个搜索词作为引理,我们使用当代德语词典来查找该引理的所有变形和派生形式。然后,我们应用转换规则(来自训练数据)来生成历史拼写变体。对于评价,我们认为,由此产生的检索质量。实验结果表明,该方法可以显著提高历史文献的检索质量。
We present a new approach for the retrieval of texts with non-standard spelling, which is important for historic texts e.g. in English or German. In this paper, we describe the overall architecture of our system, followed by its evaluation. Given a search term as lemma, we use a dictionary of contemporary German for finding all inflected and derived forms of the lemma. Then we apply transformation rules (derived from training data) for generating historic spelling variants. For the evaluation, we regard the resulting retrieval quality. The experimental results show that we can improve the retrieval quality for historic collections substantially.