Abstract: In an era where a significant portion of interpersonal communication and self-expression transpires online, social media platforms have become rich sources of data on individual psychological states. The research focuses on developing methods to analyse these digital footprints to identify potential signs of mental disorders. Utilizing advanced Natural Language Processing (NLP) techniques and convolutional neural network (CNN) algorithms, we analyse textual content from social networks to detect patterns and markers indicative of mental health issues. We address critical ethical considerations, including user privacy, data security, and the implications of diagnostic accuracy. The findings illustrate the potential of social media mining in providing valuable insights for early mental health intervention. The system distinguishes between toxic and non-toxic words to assess the emotional and psychological well-being of users, thereby enabling more precise and meaningful analysis.
Keywords: Toxic, Non-Toxic comments, Convolutional Neural Network (CNN) text analysis, Natural Language Processing (NLP).
|
DOI:
10.17148/IJARCCE.2024.13823
