A DEEP LEARNING APPROACH TO EARLY DETECTION OF DEPRESSION THROUGH SOCIAL MEDIA TEXT ANALYSIS
Abstract
In this investigation social media text analysis will be used to utilize a model BERT-BiLSTM to find early signs of depression. The study is trying to figure out if only certain things said on social media can signal certain things going on in a persons mental health just by what they say. Many versions of pre-trained models. Online writing on social media like Facebook and more is used to give contextual clues to detect harmful attitudes in you. According to the results,83%of people who used the model showed improvement in there mental health issues, or at least certain symptoms. besides the specified method there was another more effective approach to machine learning they also wellmauch with both accuracy and corerct recall. The model is performing well but there's still a lot more to be fixed to improve its overall operations. This paper provides a cost-effective way to determine whether a person has depression at the very begining. As a result, their health gets better due to having better help.
Keywords : Depression detection, deep learning, BERT, BiLSTM, social media, natural language processing, sentiment analysis, machine learning, mental health, early intervention.
