Research on Sarcasm Detection of News Headlines Based on Bert-LSTM
Hongying Liu, Ling Xie · 2021 IEEE International Conference on Emergency Science and Information Technology (ICESIT) · 2021
Sarcasm detection is usually difficult to be identified. Most of the existing sarcasm detection is based on the content of social networking sites, and there is very little sarcasm detection of news headlines. This paper proposes a news headline sarcasm detection based on Bert LSTM model, using Bert to process the word vector, taking the output result as the input of LSTM. The news headline data set written by public professionals which is deemed as professional and low noise, is used to verify this model, resulting that the model has a good recognition rate.