PCD (Poem Comprehensive Dataset)

“阿拉伯语数据集主要从 الموسوعة الشعرية 和 الشعرية 和 الديوان 抓取。合并两者后，诗句总数为 1,831,770 诗句。每节诗都标有韵律、作者和时代写在。有22米，3701位诗人和11个时代：前伊斯兰，伊斯兰，倭马亚，马穆鲁克，阿拔斯，阿尤布，奥斯曼，安达卢西亚，倭马亚和阿拔斯之间的时代，法蒂玛，最后是现代。我们只感兴趣在归因于 Al-Farahidi 的 16 个经典米中，它们构成了数据集的大部分，总数约为 170 万节经文。需要注意的是，经文的变音符号状态并不一致。这意味着一个经文可以携带完整的，半变音符号，或者什么都没有。英语数据集是从许多不同的网络资源中抓取的。它由 199,002 节经文组成，每节都标有以下四个米之一：Iambic、Trochee、Dactyl 和 Anapaestic。 Iambic 类主导数据集吨；它们是 186,809 首 Iambic 诗句、5418 首 Trochee 诗句、5378 首 Anapaestic 诗句、1397 首 Dactyl 诗句。”

The Arabic dataset was primarily crawled from *Al-Mawsu‘ah al-Shi‘riyyah* and *Al-Diwan*. After merging the two sources, the total number of poetic verses reaches 1,831,770. Each verse is annotated with its meter, poet, and historical period. The dataset covers 22 poetic meters, 3701 poets, and 11 historical periods: Pre-Islamic, Islamic, Umayyad, Mamluk, Abbasid, Ayyubid, Ottoman, Andalusi, the transitional period between Umayyad and Abbasid eras, Fatimid, and Modern. We only focus on the 16 classic meters attributed to Al-Farahidi, which make up the majority of the dataset, with a total of approximately 1.7 million verses. It should be noted that the diacritic status of the verses is inconsistent. This means that a verse may have full diacritics, partial diacritics, or no diacritics at all. The English dataset was crawled from a variety of different web resources. It consists of 199,002 poetic verses, each annotated with one of the following four meters: Iambic, Trochee, Dactyl, and Anapaestic. The Iambic class dominates the dataset: there are 186,809 Iambic verses, 5418 Trochee verses, 5378 Anapaestic verses, and 1397 Dactyl verses.

搜集汇总

背景与挑战

背景概述

PCD（Poem Comprehensive Dataset）是一个综合诗歌数据集，专注于诗韵分类，包含阿拉伯语和英语诗歌。阿拉伯语部分拥有约183万诗句，标注了韵律、作者和时代等信息；英语部分包含约19.9万诗句，分为四种韵律类型。该数据集旨在支持语言理解和合成研究。