Models developed by poltextLAB were applied in new research on sentiment analysis in political communication. The article by Kamran Aziz, Donghong Ji, Prasun Chakrabarti, Tulika Chakrabarti, Muhammad Shahid Iqbal, and Rashid Abbasi, titled Unifying aspect-based sentiment analysis BERT and multi-layered graph convolutional networks for comprehensive sentiment dissection, was published in Scientific Reports. The study examines how combining BERT-based models with graph convolutional networks can provide a more nuanced analysis of sentiment dimensions. The study is available here: https://doi.org/10.1038/s41598-024-14646

Researchers and practitioners can also access the models developed by poltextLAB on HuggingFace: https://huggingface.co/poltextlab