Wiki Dumps to Training Corpora: South Slavic Case

๐Ÿ“… 2026-04-28
๐Ÿ“ˆ Citations: 0
โœจ Influential: 0
๐Ÿ“„ PDF
๐Ÿค– AI Summary
This work proposes a systematic methodology for constructing high-quality training corpora for South Slavic language models from raw Wikimedia data. Starting with multilingual Wiki project texts, the approach first parses Wiki markup to extract natural language content and then employs an n-gramโ€“based redundancy detection mechanism to effectively filter out highly repetitive, low-information articles. This pipeline substantially enhances the linguistic richness and authenticity of the resulting corpus while maintaining cross-lingual applicability. The final resource encompasses seven South Slavic languages, offering a reliable foundation for large language model training and cross-linguistic comparative studies.
๐Ÿ“ Abstract
This paper presents a methodology for transforming raw Wikimedia dumps into quality textual corpora for seven South Slavic languages. The work is divided into two major phases. The first involves extracting and cleaning text from raw dumps of Wikipedia, Wikisource, Wikibooks, Wikinews, and Wikiquote, where available. This step requires careful handling of raw wiki markup to isolate, first of all, textual articles, and then usable natural language text within them. The second phase addresses the challenge of suspicious or low-quality articles, which are often generated from databases or structured knowledge bases. These articles are characterised by repetitive patterns, generic phrasing, and minimal to no original content. To mitigate their impact, a n-gram-based filtering strategy was employed to detect high levels of textual redundancy between articles and then remove such articles from the corpora entirely. The resulting datasets aim to provide linguistically rich texts suitable for training language models or conducting comparative research across South Slavic languages. By combining systematic extraction with quality control, this work contributes to the creation of reliable, high-information corpora that reflect authentic language use and cultural context. While focused on the South Slavic case in the paper, the approach is mostly language-agnostic and can be generalised to other languages and language families.
Problem

Research questions and friction points this paper is trying to address.

Wikimedia dumps
textual corpora
South Slavic languages
low-quality articles
textual redundancy
Innovation

Methods, ideas, or system contributions that make the work stand out.

n-gram filtering
wiki dump processing
low-quality article detection
textual redundancy removal
multilingual corpus construction
๐Ÿ”Ž Similar Papers
No similar papers found.
๐Ÿ’ผ Related Jobs
No related jobs found.
M
Mihailo ล koriฤ‡
University of Belgrade, Faculty of Mining and Geology