Mood Lens: AI-Driven Sentiment Detection And Mental Health Monitoring From Social Platforms

Abstract

This paper presents an advanced computational system designed to assess sentiment patterns and analyze mental health trends related to depression and anxiety. The increase in microblogging and social media updates has increased companies to public sentiment extraction practices. We want to extract information from both the sentiments we analyze and a vast amount of data that will classify the user perspective for a positive or negative sentiment. Further, we would like to improve classification by using hashtags as an additional means of filtering results. We will use ML algorithms, such as XGBoost, SVC and Random Forest to train the machine learning models to detect sentiments and then determine which model works best, based on accuracy. Finally, we would like to create a chatbot, using NLP, to determine sentiments with the users, while maintaining user and bot confidentiality. This will allow individuals to gain an understanding of their mental state, while obtaining a proper line of action for self-regulation. All this will require secured, end-to-end encryption due to the sensitive nature of the data to keep chats between him/herself and the bot completely private and secure.