roberta-base-suicide-prediction-phr
| Property | Value |
|---|---|
| Base Model | RoBERTa-base |
| License | MIT |
| Language | English |
| Accuracy | 96.53% |
| F1 Score | 96.52% |
What is roberta-base-suicide-prediction-phr?
This is a specialized fine-tuned version of RoBERTa-base designed to detect suicidal tendencies in text content. The model has been trained on a carefully curated dataset from Reddit, achieving remarkable performance metrics with 96.53% accuracy. It represents a significant advancement in automated mental health monitoring tools.
Implementation Details
The model utilizes the robust RoBERTa architecture and was trained on an RTXA5000 GPU using carefully selected hyperparameters. The training process involved a dataset of approximately 186,000 samples for training and 23,000 for evaluation, split in an 80:10:10 ratio.
- Learning rate: 2e-05 with Adam optimizer
- Batch size: 16 for both training and evaluation
- Training duration: 3 epochs
- Extensive data preprocessing including lowercase conversion, special character removal, and lemmatization
Core Capabilities
- Binary classification of text for suicidal tendencies
- High precision (96.38%) and recall (96.66%)
- Robust performance on cleaned and preprocessed text
- Efficient processing of social media content
Frequently Asked Questions
Q: What makes this model unique?
The model combines state-of-the-art RoBERTa architecture with extensive data preprocessing and achieves exceptional performance metrics in suicide risk detection, making it particularly valuable for mental health monitoring applications.
Q: What are the recommended use cases?
The model is best suited for analyzing text content for potential suicide risk indicators, particularly in social media monitoring, mental health support systems, and research applications. However, it should be used as part of a comprehensive assessment system and not as a standalone diagnostic tool.
