This paper details our classification model developed for the 10th Social Media Mining for Health Research and Applications Workshop (SMM4H 2025), addressing Task 6 focused on a binary classification task to distinguish Reddit posts that contain mentions of adverse vaccine events. Our objective was to enhance the detection of tweets discussing adverse reactions to herpes zoster (shingles) vaccines from other vaccine-related discussions. To do this, we used a pre-trained language model, RoBERTa in various sizes and various training methods. As we observe unstable fluctuations in performance metrics during training, we implement the ensemble approach of bagging that combines predictions from different models. Although our best-performing model achieved F1 scores of 0.96 on the validation set and 0.94 on the test set, the experiment indicates that the bagging approach contributes to improved generalizability.

PDF