2024 NAVER Cloud Platform Capstone Project. An AI-powered chatbot built on NCP services that analyzes financial statements and provides insights through natural language interactions.
FinBot-NCP is a financial analysis chatbot that leverages RAPTOR's tree-structured retrieval methodology and extends it with metadata-based clustering for enhanced financial document analysis. This project builds upon the RAPTOR framework while incorporating ideas from FinRAG to better handle financial data.
- Tree-structured retrieval for financial documents
- Metadata-based clustering (by sector, company, year)
- Natural language processing of financial statements
- Integration with NCP services
- Interactive financial analysis capabilities
git clone https://github.com/YourUsername/finbot-ncp.git
cd finbot-ncp
pip install -r requirements.txt# Example code will be added as development progresses- NCP configuration steps
- API key setup
- Service integration details
finbot-ncp/
├── src/ # Source code
├── tests/ # Test files
├── data/ # Data processing scripts
├── models/ # Model implementations
└── docs/ # Documentation
- RAPTOR integration for financial documents
- FinQA dataset preprocessing
- Metadata-based clustering implementation
- NCP service integration
- Financial analysis capabilities
- Interactive chat interface
This project builds upon the following works:
-
RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval
- Original Repository
- Paper
- Authors: Sarthi, P., Abdullah, S., Tuli, A., Khanna, S., Goldie, A., & Manning, C. D.
-
FinRAG: A Retrieval-Based Financial Analyst
- Authors: Chawla, K., & Naliath, A.
- Department of Computer Science, Stanford University
Contributions are welcome! Please feel free to submit a Pull Request.
This project is licensed under the MIT License - see the LICENSE file for details.
- NAVER Cloud Platform
- Original RAPTOR team
- Stanford CS Research Team behind FinRAG