A multi-label text sentiment analysis model based on sentiment correlation modeling
2024

ECO-SAM: A New Model for Sentiment Analysis

Sample size: 32445 publication Evidence: high

Author Information

Author(s): Ni Yingying, Ni Wei

Primary Institution: Shanghai Jiao Tong University

Hypothesis

Can an emotion correlation-enhanced sentiment analysis model improve the accuracy of text emotion recognition?

Conclusion

The ECO-SAM effectively models sentiment semantics and achieves excellent classification performance in public sentiment analysis datasets.

Supporting Evidence

  • The ECO-SAM improved precision by 13.33%, recall by 3.69%, and F1 score by 8.44% compared to baseline models.
  • The model demonstrated effective emotion correlation modeling through extensive experiments.
  • Results showed that emotions with similar connotations have strong semantic correlations.

Takeaway

This study created a new model that helps computers understand emotions in text better, making it easier to tell how people feel based on what they write.

Methodology

The ECO-SAM uses a pre-trained BERT encoder and a self-attention mechanism to analyze text emotions.

Limitations

The model is limited to text-only data and lacks large-scale high-quality training data.

Statistical Information

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.3389/fpsyg.2024.1490796

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication