Data Quality Matters: Suicide Intention Detection on Social Media Posts Using RoBERTa-CNN

Keywords:

Suicide Intention Detection, Natural Language Processing, RoBERTa, CNN, Data Quality, OpenAI API.

Contributions:

• We developed a RoBERTa-CNN model to do SID on the SDD dataset, an online corpus dataset.
• The proposed automatic SID system achieved a mean accuracy of 98%, higher than that in the state-of-the-art methods on the same dataset.
• We find the best text embedding length for the SDD dataset by fine-tuning the relative hyperparameter.
• We used OpenAI API to clean up the dataset due to the key of data quality in training RoBERTa-CNN.

Production: Apply RoBERTa-CNN to do SID

Used Framework:

BibTex:

@INPROCEEDINGS{10782647,
  author={Lin, Emily and Sun, Jian and Chen, Hsingyu and Mahoor, Mohammad H.},
  booktitle={2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)}, 
  title={Data Quality Matters: Suicide Intention Detection on Social Media Posts Using RoBERTa-CNN}, 
  year={2024},
  volume={},
  number={},
  pages={1-5},
  keywords={Training;Visualization;Social networking (online);Data integrity;Biological system modeling;Semantics;Data models;Robustness;Convolutional neural networks;Standards;Suicide Intention Detection;NLP;RoBERTa;CNN;Data Quality;OpenAI API},
  doi={10.1109/EMBC53108.2024.10782647}}

Recommended citation: E. Lin, J. Sun, H. Chen and M. H. Mahoor, "Data Quality Matters: Suicide Intention Detection on Social Media Posts Using RoBERTa-CNN," 2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Orlando, FL, USA, 2024, pp. 1-5, doi: 10.1109/EMBC53108.2024.10782647.
Download Paper