Detection of Author’s Personalities: Application to Identifying Psychological Disorders on Social Media Based on Transformers

The proliferation of social media offers a unique opportunity to analyze user behaviors and personality traits. This paper explores the application of deep learning for detecting author personality traits, with a specific focus on identifying psychological disorders. We present a methodology that integrates advanced natural language processing (NLP) techniques and deep neural networks to analyze texts posted on social media. Using annotated datasets, we train and evaluate various deep learning models to predict indicators of psychological disorders such as depression, anxiety, and bipolar disorder. Our results demonstrate that deep learning approaches outperform traditional text classification methods, offering promising tools for large-scale mental health monitoring. This study highlights the ethical and practical implications of automated psychological disorder detection, while proposing ways to improve the accuracy and reliability of predictive models.