TRESP: Modeling Group-level Public Sentiment in Social Networks through Topic and Role Enhancement

Overview

TRESP is a group-level public sentiment prediction model, that predicts emotional scores for positive, negative, and neutral emotions. Additionally, the prediction range can be refined by adjusting the size of the time window.

Dataset

We evaluate our method on the TwiCovid19 and PHEME dataset.

Quick Start

Step1: Model Training

python train.py

Step2: Model Testing

python test.py

Citation

Please cite our repository if you use TRESP in your work.

@article{zhang2024modeling,
  title={Modeling group-level public sentiment in social networks through topic and role enhancement},
  author={Zhang, Ruwen and Liu, Bo and Cao, Jiuxin and Zhao, Hantao and Sun, Xuheng and Liu, Yan and Sun, Xiangguo},
  journal={Knowledge-Based Systems},
  pages={112594},
  year={2024},
  publisher={Elsevier}
}