Exploring Machine Learning Techniques for Irony Detection

抄録

Irony detection is considered a complex task in Natural Language Processing. This paper first introduce and cover the recently state of irony detection. Then we review and summarize previous related research on text-based irony detection. Finally we compare various classifiers including the proposed CNN model on three dataset of tweets, and analysis and discuss the results. We conclude that CNN is effective for irony detection under various situation with our model outperforming all the other classifiers.