
Arabic Sentiment Analysis dataset
Version 1.0
17 March 2015
Copyright (C) 2015 National Research Council Canada (NRC)
Contact: Mohammad Salameh (msalameh@ualberta.ca)
	 Saif Mohammad (saif.mohammad@nrc-cnrc.gc.ca)		 
	 Svetlana Kiritchenko (Svetlana.Kiritchenko@nrc-cnrc.gc.ca)

Terms of use:
1. This dataset can be used freely for research purposes. 
2. The papers listed below provide details of the creation and use of 
   the dataset. If you use a dataset, then please cite the associated 
   papers.
3. If interested in commercial use of the dataset, send email to the 
   contact. 
4. If you use the dataset in a product or application, then please 
   credit the authors and NRC appropriately. Also, if you send us an 
   email, we will be thrilled to know about how you have used the 
   dataset.
5. National Research Council Canada (NRC) disclaims any responsibility 
   for the use of the dataset and does not provide technical support. 
   However, the contact listed above will be happy to respond to 
   queries and clarifications.
6. Rather than redistributing the data, please direct interested 
   parties to this page:
   [ADD LINK]

Please feel free to send us an email:
- with feedback regarding the datasets. 
- with information on how you have used the dataset. 
- if interested in a collaborative research project.

.......................................................................

Arabic Sentiment Analysis dataset
----------------------------------
The package has two datasets :

1)BBN dataset: A random subset of 1200 Levantine dialectal sentences 
  chosen from the BBN Arabic-Dialect/English Parallel Text. The sentences
  are extracted social media posts and provided with their translation.
  https://catalog.ldc.upenn.edu/LDC2012T09
  We manually annotated this subset and its translations (both manual 
  and automatic) for sentiment (positive, negative, or neutral).

2)Syrian tweets: dataset of 2000 tweets originating from Syria 
  (a country where Levantine dialectal Arabic is commonly spoken).
  These tweets were collected in May 2014 by polling the Twitter API.
  This dataset is not provided with manual English translation.
  We manually annotated this subset and its translations (both manual 
  and automatic) for sentiment (positive, negative, or neutral).
.......................................................................

PUBLICATIONS
------------
Details of the datasets can be found in the following peer-reviewed
publications:

--Sentiment After Translation: A Case-Study on Arabic Social Media Posts.
 Mohammad Salameh, Saif M Mohammad and Svetlana Kiritchenko, In 
 Proceedings of the North American Chapter of the Association for 
 Computational Linguistics (NAACL-2015), June 2015, Denver, Colorado.

Links to the papers are available here:

.......................................................................

VERSION INFORMATION
-------------------
Version 1.0 is the first version as of 17 March 2015.

.......................................................................

FORMAT
------
The BBN and Syrian dataset has sheets with these title

1)ar_manual.sent.: The manual sentiment annotation for the Arabic posts
2)en_auto.trans-manl.sent : has the automatic translation that are 
  manually annotated for sentiment
3)en_manl.trans-manl.sent: (ONLY for BBN dataset) has the manual 
  transaltion for the Arabic posts, mannually annotated for sentiment 

Annotation categories are: positive, negative, neutral and both.
The above datasets are also provided with the confidence of the 
annotation calculated by CrowdFlower.

.......................................................................
