malek-hedhli/Arabic-Dialects-Dataset

Arabic tweets, posts and comments in 5 Arabic dialects represent as follows 4834 Tunisian 4834 MGH 4832 GLF 4832 LEV 4830 EGY The dataset have 2 columns "Sentence" and "Dialect" # Arabic-Dialects-Dataset Arabic tweets, posts and comments in 5 Arabic dialects represent as follows : 4834 Tunisian 4834 MGH 4832 GLF 4832 LEV 4830 EGY The dataset have 2 columns "Sentence" and "Dialect" These data are presented in these 2 .xlsx files the first "CleanedDataset.xlsx" file contains the following pre-processing : -Rmove "arabic" Stop Words -Remove hashtag -Remove punctuation -Normalize -Remove repeating char -Stemming -Replace @username with empty string -Remove Links -Balance dialect numbers -Tokenize the Sentences and the second "CleanedWithoutTokenizer.xlsx" contains the same cleaning and preprocessing procedure without removing stop words and without tokenize