Chinese Metaphor Sentiment Analysis Based on Attention-Based LSTM

The metaphor is a kind of frequently used figurative language. Metaphor computing is helpful in many natural language processing tasks, such as opinion mining, discourse understanding, and dialogue system. This article presents a Chinese metaphor sentiment analysis approach based on an attention-based Long Short-Term Memory (LSTM) network. Our hypothesis is that the target and the context are important in metaphor sentiment analysis, and their interaction provides reliable sentiment-related classification features. The experimental results agree with our hypothesis and justify the effectiveness of the model.