TY - JOUR
T1 - IMPROVED DEEP LEARNING SENTIMENT ANALYSIS FOR ARABIC
AU - Binmahfoudh, Ahmed
N1 - KAUST Repository Item: Exported on 2023-05-24
Acknowledgements: The author is grateful to help from Dr. Seifeddine Mechti (University of Sfax) as one of the team. Also, to King Abdullah University of Science and Technology (KAUST) in performing this competition and providing the data.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.
PY - 2023/2/15
Y1 - 2023/2/15
N2 - Sentiment Analysis (SA) has recently gained great interest in Natural Language Processing (NLP). In fact, NLP consists in extracting data from texts and categorizing certain tweets as Positive, Negative, or Neutral. In this paper, we also present our participation in the Arabic Sentiment Analysis Challenge organized by King Abdullah University of Science and Technology (KAUST). Data of interest are tweets written in Arabic language, which becomes more challengeable. In this manuscript, we present the introduced system and the bi-LSTM model. Also, detail the less efficient explored solutions. Our main objective is to extract the crucial semantic data in Arabic tweets. The obtained findings about Arabic twitter corpus reveal that the performance of the developed technique is better than that proposed in the literature. Official test accuracy scores are 0.7605 with Macro-F1 score.
AB - Sentiment Analysis (SA) has recently gained great interest in Natural Language Processing (NLP). In fact, NLP consists in extracting data from texts and categorizing certain tweets as Positive, Negative, or Neutral. In this paper, we also present our participation in the Arabic Sentiment Analysis Challenge organized by King Abdullah University of Science and Technology (KAUST). Data of interest are tweets written in Arabic language, which becomes more challengeable. In this manuscript, we present the introduced system and the bi-LSTM model. Also, detail the less efficient explored solutions. Our main objective is to extract the crucial semantic data in Arabic tweets. The obtained findings about Arabic twitter corpus reveal that the performance of the developed technique is better than that proposed in the literature. Official test accuracy scores are 0.7605 with Macro-F1 score.
UR - http://hdl.handle.net/10754/691987
UR - http://www.jatit.org/volumes/hundredone3.php
UR - http://www.scopus.com/inward/record.url?scp=85153799141&partnerID=8YFLogxK
M3 - Article
SN - 1992-8645
VL - 101
SP - 1251
EP - 1260
JO - Journal of Theoretical and Applied Information Technology
JF - Journal of Theoretical and Applied Information Technology
IS - 3
ER -