Emotion recognition of Arabic tweets on Saudi Arabia social issue
○Ghadeer Attar,杉本 徹(芝浦工大)
Language is a very powerful tool humans have to communicate with one another. With communication comes emotions and it can be expressed in different ways one being through text. Nowadays, the focus is on textual communication, specifically the usage of social media, such as twitter. People uses twitter to discuss serious matters such as social and political issues. However, the data is unstructured and it requires different steps of preprocessing and classifying in order to get information out of it.
This paper introduces a technique of emotions recognition from Arabic text in a Saudi dialect regarding a social issue. Keyword spotting technique was created which takes a tweet as an input and then output is an emotion class. Starting from collecting the data from Twitter, preprocessing it and then tokenization was performed. Emotion keywords were identified for the main step which is keyword matching and extract the emotion class associated with the text.

footer 著作権について 倫理綱領 プライバシーポリシー セキュリティ 情報処理学会