A Study of Using Language Models to Detect Sarcasm

摘要

Organizations that provide online services solicit users to provide feedback to the services that their users receive. This feedback enables the organization to gauge user-satisfaction. To automatically process users’ feedback, many organizations have used natural language processing and machine learning techniques to analyze users’ sentiment from their feedback. In recent years, many language models have been developed for carrying out sentiment analysis. In this paper, we evaluate the effectiveness of several recently developed language models for carrying out sentiment analysis. The result of this research should help organizations in choosing an effective language model for analyzing feedback to monitor user-satisfaction rate.