Elamur Research⌕
☰
home / papers / 2015 / A Light Lexicon-based Mobile Application for Sentiment Mining of Arabic Tweets

A Light Lexicon-based Mobile Application for Sentiment Mining of Arabic Tweets

Paper · 2015 · Gilbert Badaro, Ramy Baly, Rana Akel et al.

Summary

The paper addresses sentiment analysis of Arabic tweets on a phone. Full Arabic NLP toolkits such as MADAMIRA cannot run on the device without a server, and server calls cost energy and response time. The authors replace heavy morphological processing with a sentiment lexicon stored locally. The lexicon is ArSenL, their own large-scale Standard Arabic resource. It was built by linking Arabic WordNet and AraMorph lemmas to English SentiWordNet, so each entry gets positive, negative and objective scores that add up to one.

To skip lemmatization, they derive a stemmed copy of ArSenL. The pipeline then tokenizes each tweet, cleans hashtags, stems words with Khoja's stemmer, and looks the stems up in the lexicon. The three score totals per tweet feed three balanced binary decision trees. Handcrafted rules merge the three outputs into positive, negative or neutral.

On a published Arabic tweet set, the full system reaches roughly 67% accuracy. The work shows that a cheap, interpretable, lexicon-driven method can deliver usable Arabic opinion mining on constrained hardware. It is released as a public app with visual topic summaries.

Key points

  • ArSenL combines English WordNet, Arabic WordNet, English SentiWordNet (ESWN) and AraMorph. Each lemma carries three scores (positive, negative, objective) summing to 1. It can be browsed at oma-project.com and downloaded for research.
  • Words missing from the stemmed lexicon receive zero for all three scores. Summing scores per tweet beat averaging them in the authors' tests.
  • Classifier accuracies:
    • positive/not positive: 61.2%
    • negative/not negative: 72.9%
    • objective/not objective: 67.8%
    • combined system: 67.3%
  • The app takes a topic, pulls recent Arabic tweets, and colours each tweet by sentiment. It also shows frequent hashtags and a dated history of topic sentiment.
  • Processing 20 tweets per request was too slow, so the app fetches batches of 10. Handling negation is left as future work.

Interesting ideas

  • I1. When input normalization must stay cheap, the authors coarsen the lexicon to the same level (stems) instead of upgrading the input processing (lemmas), so lookup keys always agree.
  • I2. Hashtags are not thrown away: the marker is stripped and the underlying word is kept, because it often carries the opinion.
  • I3. Polarity scores attached to English synsets are carried over to another language through WordNet links, giving a large lexicon without new annotation.
  • I4. Rather than one three-way classifier, the authors train separate balanced yes/no detectors per label and resolve conflicts with explicit rules, which keeps each decision inspectable.
  • I5. The authors report accuracy per component as well as for the whole system, and test aggregation choices (sum vs average) explicitly.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (tokens): keep a collapsed-key fallback index for lexicon lookups, used only when the lemmatizer is unsure, and log which key level produced the match.
  • I2 (tokens): the tokenizer can treat # and similar social markup as a removable wrapper and lemmatize and tag the inner word normally, keeping the original span for explanation.
  • I3 (world): values or evaluative properties attached to WordNet synsets can be propagated through synset IDs into the ontology and knowledge graph, keeping provenance per entry.
  • I4 (frames): detectors for irony, speech-act type or evaluative stance could be independent binary judgments over frames, with a logged rule step that explains the final label.
  • I5 (learn): evaluations can always report each sub-decision separately and record ablations of aggregation operators, so regressions point to a specific stage.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓American University of BeirutBeirut · Lebanon34 works · 2,561 citations🎓Qatar UniversityDoha · Qatar48 works · 1,026 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 2Research subjects and languages the work is about.Sentiment & emotion · 972 worksSemitic languages · 2,908 works

Cite

Cite as Gilbert Badaro, Ramy Baly, Rana Akel, Linda Fayad, Jeffrey Khairallah, Hazem Hajj, Khaled Shaban and Wassim El-Hajj. 2015. A Light Lexicon-based Mobile Application for Sentiment Mining of Arabic Tweets. Proceedings of the Second Workshop on Arabic Natural Language Processing.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/w15-3203

@inproceedings{badaro2015light,
  title = {{A Light Lexicon-based Mobile Application for Sentiment Mining of Arabic Tweets}},
  author = {Gilbert Badaro, Ramy Baly, Rana Akel, Linda Fayad, Jeffrey Khairallah, Hazem Hajj, Khaled Shaban and Wassim El-Hajj},
  booktitle = {Proceedings of the Second Workshop on Arabic Natural Language Processing},
  year = {2015},
  doi = {10.18653/v1/w15-3203},
  url = {https://doi.org/10.18653/v1/w15-3203}
}

References

26 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Abdul-Mageed, M., Diab, M. and Korayem, M. (2011). Subjectivity and sentiment analysis of modern standard Arabic. In Proceedings of the 49 th Annual Meeting of the Association for Computation Linguistics: Human Language Technologies: short papers- Volume2. Association for Computational Linguistics.
  2. Abdul-Mageed, M., & Diab, M. (2012). Toward building a large-scale Arabic sentiment lexicon. In Proceedings of the 6 th International Global Word-Net Conference (pp. 18-22). find
  3. Abdul-Mageed, M., & Diab, M. (2014). SANA: A Large Scale Multi-Genre, Multi-Dialect Lexicon for Arabic Subjectivity and Sentiment Analysis. In Proceedings of the Language Resources and Evaluation Conference (LREC). Reykjavik, Iceland.
  4. Alhazmi, S., Black, W., & McNaught, J. (2013). Arabic SentiWordNet in Relation to SentiWordNet 3.0. 2180-1266, 4(1), 1-11. find
  5. Aly, M. A., & Atiya, A. F. (2013, August). LABR: A Large Scale Arabic Book Reviews Dataset. In ACL (2) (pp. 494-498). ↗
  6. AraMorph. (n.d.). Retrieved May 19, 2015, from http://sourceforge.net/projects/aramorph/ ↗
  7. Baccianella, S., Esuli, A., & Sebastiani, F. (2010, May). SentiWordNet 3.0: An Enhanced Lexical Resource for Sentiment Analysis and Opinion Mining. In LREC (Vol. 10, pp. 2200-2204).
  8. Badaro, G., Baly, R., Hajj, H., Habash, N., & El-Hajj, W. (2014). A Large Scale Arabic Sentiment Lexicon for Arabic Opinion Mining. ANLP 2014, 165.
  9. Black, W., Elkateb, S., & Vossen, P. (2006). Introducing the Arabic wordnet project. In Proceedings of the third International WordNet Conference (GWC-06). find
  10. Davidov, D., Tsur, O., & Rappoport, A. (2010, August). Enhanced sentiment learning using twitter hashtags and smileys. In Proceedings of the 23rd International Conference on Computational Linguistics: Posters (pp. 241-249). Association for Computational Linguistics.
  11. Denecke, K. (2008, April). Using sentiwordnet for multilingual sentiment analysis. In Data Engineering Workshop, 2008. ICDEW 2008. IEEE 24 th International Conference on (pp. 507-512). find
  12. IEEE Elarnaoty, M., AbdelRahman, S., & Fahmy, A. (2012). A Machine Learning Approach for Opinion Holder Extraction in Arabic Language. International Journal of Artificial Intelligence & Applications, 3(2). ↗
  13. El-Beltagy, S. R., & Ali, A. (2013, March). Open issues in the sentiment analysis of Arabic social media: A case study. In Innovations in Information Technology (IIT), 2013 9th International Conference on (pp. 215-220). IEEE. find
  14. El-Halees, A. (2011). Arabic opinion mining using combined classification approach. find
  15. Esuli, A., & Sebastiani, F. (2005). Determining the semantic orientation of terms through gloss classification. In Proceedings of the 14 th ACM international conference on Information and knowledge management (pp. 617-624). ACM. find
  16. Esuli, A., & Sebastiani, F. (2006). Sentiwordnet: A publicly available lexical resource for opinion mining. In Proceedings of LREC (Vol. 6. pp. 417-422).
  17. Feldman, R. (2013). Techniques and applications for sentiment analysis. Communications of the ACM, 56(4), 82-89. ↗
  18. Graff, D., Maamouri, M., Bouziri, B., Krouna, S., Kulick, S., & Buckwalter, T. (2009). Standard Arabic Morphological Analyzer (SAMA) Version 3.1, 2009. Linguistic Data Consortium LDC2009E73. Habash, Nizar. Introduction to Arabic Natural Language Processing, Synthesis Lectures on Human Language Technologies, Graeme Hirst, editor. find
  19. Morgan & Claypool Publishers., 2010. find
  20. Khoja, S., & Garside, R. (1999). Stemming arabic text. Lancaster, UK, Computing Department, Lancaster University. find
  21. Mourad, A., & Darwish, K. (2013, June). Subjectivity and sentiment analysis of modern standard Arabic and Arabic microblogs. In Proceedings of the 4th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis (pp. 55-64).
  22. Mukherjee, S., Malu, A., AR, B., & Bhattacharyya, P. (2012, October). TwiSent: a multistage system for analyzing sentiment in twitter. In Proceedings of the 21st ACM international conference on Information and knowledge management (pp. 2531-2534). ACM. find
  23. Ohana, B., & Tierney, B. (2009, October). Sentiment Classification of reviews using SentiWordNet. In 9 th . IT & T Conference (p.13). find
  24. Pasha, Arfath, Mohamed Al-Badrashiny, Mona Diab, Ahmed El Kholy, Ramy Eskander, Nizar Habash, Manoj Pooleery, Owen Rambow and Ryan M. Roth. MADAMIRA: A Fast, Comprehensive Tool for Morphological Analysis and Disambiguation of Arabic. In Proceedings of the Language Resources and Evaluation Conference (LREC), Reykjavik, Iceland, 2014.
  25. Radcliffe, D. (2013). Twitter takes off in Saudi – and other news of social media in the Arab world. British Broadcasting Corporation. find
  26. Taboada, M., Brooke, J., Tofiloski, M., Voll, K., & Stede, M. (2011). Lexicon-based methods for sentiment analysis. Computational linguistics, 37(2), 267-307. ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/badaro-2015-light-lexicon-based-mobile-application/