Vaccine hesitancy in Türkiye: A natural language processing study on social media

Vaccine hesitancy is a significant public healthcare problem that is threatening everyone worldwide. Vaccine hesitancy has become more ingrained in Turkish society, mainly through social media. Unfortunately, reflections of this hesitancy are preventable deaths or permanent disabilities. Because of the uncontrolled spread of misinformation and disinformation on social media, Türkiye is facing a future health crisis. As a step towards preventing this crisis, our main objective is to use the power of artificial intelligence techniques on Turkish social media posts to detect antivaccine posts. Through this study, it will be possible to raise awareness about the importance of vaccines in Turkish society, strengthen Türkiye’s defense mechanism against potential epidemics, and ensure that our society exchanges information in a healthier digital environment. We collected and cleaned a novel Turkish social media dataset, resulting in 3778 posts. Then, we used a baseline machine learning method, logistic regression, popular machine learning methods, support vector machines, and XGBoost to detect antivaccine thoughts and misleading information from Turkish social media posts. Further, we included transformers that changed the natural language processing domain. Evaluations are conducted using a multilingual BERT and two models specifically trained for recognizing Turkish texts, such as BERTurk. Results showed that transformers can separate Turkish social media posts with antivaccine beliefs from other posts with a 75.9% Area Under the ROC curve rate.