Emotion Detection in Arabic Texts Extracted from Twitter Network by Using Machine Learning Techniques

https://doi.org/10.55529/jaimlnn.21.27.36

Authors

  • Hiba Mohamed Alkhateeb Master in Web Science, Syrian Virtual University, Damascus, Syria
  • Ammar Alnahhas Ph.D in Syrian Virtual University, Damascus University

Keywords:

Arabic Text, Emotion Detection, Twitter, Social Media, Xgboost Algorithm,

Abstract

Arabic sentiment analysis research existing currently is very limited. While sentiment analysis has many applications in English, the Arabic language is still recognizing its early steps in this field. In this paper, we show an application on Arabic sentiment analysis by implementing a sentiment classification for Arabic tweets. The retrieved tweets are analyzed to provide their sentiments polarity (positive, or negative). Since, this data is collected from the social network Twitter; it has its importance for the Middle East region, which mostly speaks Arabic our research was characterized by relying mainly on the XGBoost algorithm, which is considered the most powerful algorithm in the world of big data analysis. The results were compared with several other algorithms that were used to show us the clear difference in terms of accuracy and F1 score between these algorithms and the XGBoost algorithm, The hyperparameters are obtained from the grid search algorithm to get the best performance for each model used in our work.
The results showed that our use of the XGBoost model gave more accurate results compared to the previous best-performing models in this field. A website has been created so that the user enters the comment to show him the result (positive comment or negative comment).

Published

2022-01-14

How to Cite

Hiba Mohamed Alkhateeb, & Ammar Alnahhas. (2022). Emotion Detection in Arabic Texts Extracted from Twitter Network by Using Machine Learning Techniques. Journal of Artificial Intelligence,Machine Learning and Neural Network , 2(01), 27–36. https://doi.org/10.55529/jaimlnn.21.27.36

Similar Articles

1 2 3 4 > >> 

You may also start an advanced similarity search for this article.