Political Arabic Article Dataset

PAAD: Political Arabic Article Dataset is a collection of political Arabic text, which covers modern Arabic language used in newspaper, blogs and social network. PAAD can be used in different Arabic NLP tasks such as Text Classification, Target, Article Orientation and Word Embedding. The text contains alphabetic, numeric, English word and symbolic words. The documents in the dataset are categorized into 3 classes: Reform "اصلاحي", Conservative "محافظ" and Revolutionary "ثوري". The number of documents for each class:
Reform = 80
Conservative = 58
Revolutionary = 68
PAAD contains a total number of 206 articles. Articles were manually collected and using python scripts specifically for Excel file. There are two Excel file first original file this file same raw data but in excel file second file with Arabic normalization:
1- إأٱآا = ا
2- ي = ى
3- ؤ ئ = ء
4- ة = ه
5- Remove diacritics as (ُ,ْ,َ,ِ,ّ,~,ً,ٍ,ٌ)
How to use it:
___________
1. Unzip compressed resources.
2. There are three main folders each folder labelled by the category's name as Reform = S, Conservative = M and Revolutionary = T.
3. Each folder contains a set of article files corresponding to its category.
4. There are 2 excel file first one as raw data but in one file with the label for each article second excel file with Arabic normalization.

V1 = original corpus
V2 = preprocessing corpus
V3 = root stemming corpus (ISRI)
V4 = light stemming corpus (ISRI)

PAAD：阿拉伯语政治文章数据集（Political Arabic Article Dataset）是一套阿拉伯语政治文本集合，涵盖报纸、博客与社交网络中使用的现代阿拉伯语。该数据集可应用于各类阿拉伯语自然语言处理（Natural Language Processing, NLP）任务，包括文本分类、目标识别、文章立场分析与词嵌入（Word Embedding）。文本内容包含字母、数字、英文单词与符号词汇。

数据集内的文档共分为三类：改革派（Reform "اصلاحي"）、保守派（Conservative "محافظ"）与革命派（Revolutionary "ثوري"）。各类文档的数量分别为：改革派80篇，保守派58篇，革命派68篇，数据集总计包含206篇文章。

文章通过人工收集，并结合专门针对Excel文件编写的Python脚本完成处理。数据集包含两份Excel文件：第一份为原始数据文件，即存储未经处理的原始数据的Excel文件；第二份为经过阿拉伯语规范化处理的文件，具体规范化规则如下：
1. 将إأٱآا统一转换为ا
2. 将ي统一转换为ى
3. 将ؤ、ئ统一转换为ء
4. 将ة统一转换为ه
5. 移除所有变音符号，包括(ُ,ْ,َ,ِ,ّ,~,ً,ٍ,ٌ)

使用方法：
1. 解压压缩资源包。
2. 数据集包含三个主文件夹，分别以对应类别的名称标识：改革派对应文件夹名为S，保守派对应M，革命派对应T。
3. 每个文件夹内包含对应类别的多篇文章文件。
4. 此外还包含两份Excel文件：第一份为整合了所有文章及其对应标签的原始数据单文件；第二份为经过阿拉伯语规范化处理的Excel文件。

版本说明：
V1为原始语料库，V2为经过预处理的语料库，V3为基于ISRI算法的词根抽取语料库，V4为基于ISRI算法的轻量词根抽取语料库。