This project focuses on analyzing the sentiments of mental health statuses using machine learning and Natural Language Processing (NLP). By leveraging fundamental NLP methods, we processed a dataset of 52,000+ mental health-related text samples, applying text cleaning, tokenization, and vectorization approaches. We trained and fine-tuned a logistic regression model, achieving a 72% F1-score in classifying text into categories such as anxiety, depression, and suicidal thoughts. This work serves as a foundation for future applications, such as real-world mental health monitoring and chatbot development.
- Sentiment analysis using Logistic Regression model
- Preprocessing of text data for enhanced accuracy
- Data visualization to understand patterns and trends
- Model training and evaluation
- Dr. Hector Santos-Villalobos, course instructor for COSC 325 - Introduction to Machine Learning (Spring 2024).
- The dataset: Sentiment Analysis for Mental Health by Suchintika Sarkar.
- Python libraries and frameworks used in the project.