Using Hybrid-Stemming Approach to Enhance Lexicon-based Sentiment Analysis in Arabic

Awwad H., ALPKOÇAK A.

International Conference on New Trends in Computing Sciences (ICTCS), Amman, Ürdün, 11 - 13 Ekim 2017, ss.229-235, (Tam Metin Bildiri) identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası:
  • Doi Numarası: 10.1109/ictcs.2017.26
  • Basıldığı Şehir: Amman
  • Basıldığı Ülke: Ürdün
  • Sayfa Sayıları: ss.229-235
  • Anahtar Kelimeler: Sentiment analysis, Lexicon-based, Arabic, Stemming, Negation, Intensification
  • Dokuz Eylül Üniversitesi Adresli: Evet

Özet

The objective of this article is to present a hybrid-stemming approach to enhance lexicon based sentiment analysis (SA) in Arabic in both sentence-level and document-level. The new approach uses root-based and light stemmers together to enhance the likelihood of finding sentiment words. We applied the proposed hybrid-stemming approach into two datasets and four lexicons, and we provided a comparison with root-based stemmer and light stemmers individually. The results we obtained demonstrate that the presented hybrid-stemming approach is more accurate than both root-based stemmers and light stemmer when they are using separately for sentiment analysis in Arabic.