Incorporating text and emoticon features in sarcasm and irony detection: A deep learning-based approach

Emotions play a crucial role in social media interactions, and sarcasm and irony often convey complex emotions that can be challenging to detect accurately. This study focuses on automating the recognition of sarcasm and irony in Chinese through various computational methods using machine learning algorithms. This study utilizes both text and emoticon/emoji feature to improve detection accuracy. The study employs a combination of linguistic rules, word embeddings, and neural network models for detecting irony in Chinese microblog comments. The results highlight the advantages of deep learning techniques and pre-trained word embeddings in enhancing the accuracy of irony detection. Furthermore, the study demonstrates the viability of the proposed models by focusing on microblog posts related to social events. These contextualized experiments show promising results and pave the way for future studies to explore other domains and data sources. Overall, this study has made significant contributions to the field of irony recognition research and laid the groundwork for more complex models in the future, with practical implications for enhancing sentiment analysis in various fields.