This project analyzes and predicts resolution times for Jira issues using machine learning techniques.
- Retrieves issue data from MongoDB database
- Categorizes resolution times into meaningful buckets
- Analyzes resolution patterns by issue attributes (components, labels, priority, etc.)
- Creates visualizations to understand resolution time distributions
- Implements multiple prediction models:
- Text-based models (TF-IDF and BERT)
- Topic models
- Distribution-based predictions
- Stacked machine learning approach
- Clone this repository
- Install dependencies:
pip install -r requirements.txt - Optional: For BERT and topic models, install additional dependencies:
pip install sentence-transformers bertopic
python main.py --project PROJECTNAME --mongo-uri "mongodb://user:password@host:port/"Edit config.py to customize:
- Output directory
- Resolution time categories
- Model parameters
- Test/train split ratio
The tool generates:
- Visualizations of resolution time distributions
- Heatmaps showing resolution patterns by different attributes
- Trained ML models for future predictions
- Analysis reports with accuracy metrics
- Comparison of different prediction approaches