Korean Twitter Emotion Classification Using Automatically Built Emotion Lexicons and Fine-Grained Features

In recent years many people have begun to express their thoughts and opinions on Twit-ter. Naturally, Twitter has become an ef-fective source to investigate people’s emo-tions for numerous applications. Classifying only positive and negative tweets has been ex-ploited in depth, whereas analyzing finer emo-tions is still a difficult task. More elaborate emotion lexicons should be developed to deal with this problem, but existing lexicon sets are mostly in English. Moreover, building such lexicons is known to be extremely labor-intensive or resource-intensive. Finer-grained features need to be taken into account when determining finer-emotions, but many exist-ing works still utilize coarse features that have been widely used in analyzing only the po-larity of emotion. In this paper, we present a method to automatically build fine-grained emotion lexicon sets and suggest features that improve the performance of machine learning based emotion classification in Korean Twitter texts.