Mental health conditions often leave linguistic traces in everyday language. This dataset contains free-form text statements resembling anonymous online confessions, each labeled with a corresponding mental health status.
The available target labels include: Anxiety, Bipolar, Depression, Normal, Personality disorder, Stress, and Suicidal.

This dataset is built for NLP-based mental health prediction, text classification research, and early warning system development. It enables a variety of ML tasks such as:

• Multi-class mental health classification
• Context-aware sentiment analysis
• Emotion and psychological pattern detection
• Transformer model benchmarking (BERT, RoBERTa, etc.)
• Clinical text mining research (non-diagnostic)

Suitable for:

✔ Students & ML beginners

✔ Researchers

✔ Healthcare AI experiments

✔ NLP and transformer benchmarks

⚠️ Disclaimer:This dataset is not a medical tool and must not be used for real clinical diagnosis. It is intended strictly for educational, research, and machine learning purposes.