BERT-based Context-Aware Emotion Recognition Model for Improving Mental Health Assessment

Objectives: In the digital age, characterized by rapid advancements in information technology, mental health has attracted increasing attention. The evolution of natural language processing (NLP) technology has created new opportunities for psychological health assessment through emotion recognition. Traditional emotion recognition methods, such as dictionary-based approaches or simple machine learning classifiers like support vector machines (SVM) and naive Bayes, often underperform in identifying emotions in complex texts, particularly within contexts related to mental health, where emotions are diverse and nuanced.