Mental health on twitter turkey: sentiment analysis with transformers

Social media platforms are widely recognized as a valuable medium for capturing various aspects of individuals' daily lives, interests, and beliefs. Addi- tionally, these platforms have evolved into hubs for exchanging health-related infor- mation. Through collecting and analyzing social media data, valuable insights and hidden patterns regarding people's behaviors, status, and more can be extracted. This study aims to leverage machine learning techniques to perform sentiment analysis and assess the mental health of Turkish Twitter users. To accomplish this, we col- lected over 25,000 tweets containing terms related to common mental health issues, including anxiety, stress, depression, suicide, and eating disorders. The collected data was then subjected to thorough analysis, and we employed a transformer-based machine learning model to automate sentiment scoring specifically for the Turkish language. By utilizing BERT, our final deep-learning classifier achieved an im- pressive accuracy of 82.6% in predicting sentiment from the collected tweets. This study underscores the effectiveness of deep learning models and transformers in addressing Turkish natural language processing tasks. The insights derived from this research have the potential to enhance mental health services by offering a better understanding of the sentiments expressed in Turkish tweets regarding mental health.