| RoBERTa model for detecting social movement core framing tasks in tweets. | |
| Multilabel classification of 3 categories: | |
| - diagnostic (problem identification and blame) | |
| - prognostic (solutions, tactics, solidarity, and counterframing) | |
| - motivational (calls to action) | |
| The model is trained on Twitter data from three issue areas: guns, LGBTQ rights, and immigration. | |
| All data is in English and written in 2018-2019. | |
| Paper: https://journalqd.org/article/view/5896 |