Project information
- Category: LLMs (Large Language Models)
- Project date: February, 2025
- Project URL: github.com/sentiment-stocks-ai
Using LLM for sentiment analyses to recommend an investment decision: Buy, Hold, or Sell
Sentiment-Stocks-AI is a project that utilizes sentiment analysis techniques on real-time news to assist in stock investment decision-making. The system collects news about a specific company, analyzes the sentiment of these news articles using language models (LLMs), and combines this analysis with historical stock price data to recommend an investment decision: Buy, Hold, or Sell.
The project is divided into two main versions:
- DeepSeek Version: Uses the DeepSeek model (via Ollama) for sentiment analysis.
- GPT Version: Uses the GPT-4 model (from OpenAI) for sentiment analysis
Both versions follow the same logic but differ in the choice of the language model, allowing a comparison of performance and costs between a proprietary LLM (GPT) and an open-source LLM (DeepSeek).
