A Deep Learning Sentiment Analysis Method based on ERNIE and Modified DPCNN
Yang Yang, Xunde Dong, Yupeng Qiang · 2023
Pre-training models such as BERT, RoBERTa, and ERNIE have gained popularity in natural language processing (NLP). Within this domain, sentiment analysis has become an important task. People express their emotional states through their comments on social media, and sentiment analysis involves using NLP techniques to parse these comments. In the study, we propose an advanced sentiment classification model that combines ERNIE with a modified DPCNN. Our model achieved a Macro-F1 score of 87.03% on the publicly available dataset EmoInt, surpassing the performance of the six baseline models, indicating that the proposed model has competitive performance.