Skip to content

Update from task 1f32917c-66d4-417c-b5e5-e50ac7ce842d#5

Open
zouyuanqing wants to merge 1 commit into
feature/updates-v1.0.0from
github项目维护建议-e842d

Hidden character warning

The head ref may contain hidden characters: "github\u9879\u76ee\u7ef4\u62a4\u5efa\u8bae-e842d"
Open

Update from task 1f32917c-66d4-417c-b5e5-e50ac7ce842d#5
zouyuanqing wants to merge 1 commit into
feature/updates-v1.0.0from
github项目维护建议-e842d

Conversation

@zouyuanqing

Copy link
Copy Markdown
Owner

This PR was created by qwen-chat coder for task 1f32917c-66d4-417c-b5e5-e50ac7ce842d.

…s Tool Selection**

**Key features implemented:**
- Added LLM-driven agent system in stock_market_prediction/agents/llm_agent.py that autonomously selects tools and controls ARIMA/LSTM parameters
- Implemented comprehensive tool suite in stock_market_prediction/tools/__init__.py including stock data fetcher, web search, ARIMA predictor, LSTM predictor, and technical analyzer
- Integrated ReAct paradigm allowing multi-step reasoning with autonomous decision making for stock analysis
- Added network search capability enabling agent to fetch external news and market information autonomously
- Enhanced ARIMA and LSTM models with LLM-controlled hyperparameters (p,d,q values for ARIMA; hidden_size, layers, lookback days for LSTM)
- Created command-line interface in main.py with interactive analysis mode and detailed result reporting
- Added proper source attribution for OpenCLAW and Hermes Agent design inspiration in documentation
- Implemented JSON-based tool calling format with structured action-input responses for reliable function execution

The implementation creates a fully autonomous stock analysis system where the LLM agent makes independent decisions about which tools to use, what parameters to set for predictive models, and how to synthesize information from multiple sources into comprehensive analysis reports.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants