Title: Feel the Heat: Emotion Detection in Arabic Social Media Content

Year of Publication: Nov - 2014
Page Numbers: 37-49
Authors: Omneya Rabie, Christian Sturm
Conference Name: The International Conference on Data Mining, Internet Computing, and Big Data (BigData2014)
- Malaysia


The automatic detection of emotions in textual parts of social media websites such as Facebook and Twitter has applications for business development, user interface design, content creation, emergency response, among others. Current research has shown that it is possible to detect emotions for English content. To our knowledge, however, there are only few attempts for Arabic content. There is neither Arabic corpus with instances labeled for emotions, nor studies to detect emotions from Arabic microblogs content. Therefore, we collected Arabic text messages from the social networking website Twitter from January/February 2011. Human annotators labeled them with the corresponding emotions. Working with that corpus, our experiments show that emotions can be automatically detected from tweets after performing Arabic language related language preprocessing steps. Our contribution consists in adding preprocessing steps that have improved the classification results by 4.4% compared to the original Khoja stemmer. In addition, we have extracted a sample word-emotion lexicon from that corpus. Our experiment demonstrates that this sample word-emotion lexicon enhances the emotion detection results by 22.27% compared to the SMO classification using the train/test option. Finally, we show that the communication style used by the writer significantly relates with the emotion expressed in the text.