A Corpus of Arabic Literature (19-20th centuries) for Stylometric Tests

<p>The dataset contains three collections of mainly literary Arabic texts from the 19th and early 20th centuries.</p> <ul> <li><code>corpus022_JurjiZaydan_Dated</code>&nbsp;is a dated corpus of 22 historical novels by Jurjī Zaydān. It is well established that Jurjī Zaydān was publishing roughly one novel per year and the dates of publication are well known, which makes this corpus a valuable material for testing chronological changes in the style of individual writers.</li> <li><code>corpus065</code>&nbsp;is a corpus of 65 books by 8 authors;</li> <li><code>corpus300</code>&nbsp;contains 300 books by 28 authors;</li> </ul> <p>Texts have been collected from&nbsp;<a href="https://www.hindawi.org/">https://www.hindawi.org/</a>; the original EPUB files have been converted into clean text files (UTF8 encoding) and Arabic orthography has been normalized in the following manner: short vowels removed; the orthography of&nbsp;<em>alif</em>&nbsp;simplified (all&nbsp;<em>alif</em>s converted into bare&nbsp;<em>alif</em>s are used);&nbsp;<em>alif maqṣūraŧ</em>s converted to&nbsp;<em>yāʾ</em>s.</p> <p>For the names of authors and the names of works, the URIs follow the naming conventions used in the OpenITI project (<a href="https://github.com/openiti">https://github.com/openiti</a>, for the most up-to-date description, see&nbsp;<a href="https://kitab-project.org/corpus-and-data">https://kitab-project.org/corpus-and-data</a>). However, for better compatibility with R&nbsp;<code>stylo</code>,&nbsp;<em>underscore</em>&nbsp;(<code>_</code>) is used to connect URIs: AUTHOR_TITLE (in the dated Jurjī Zaydān corpus files are named YEAR_TITLE).</p>