基於 langchainai 開發，用 7 個 AI Agent 分工來模擬的避險基金（hedge fund) ：
- 市場資料 Agent - 收集和預先處理市場資料
- 估值 Agent - 計算股票的內在價值，分析公司獲利能力、成長性、財務健康狀況和估值並產生交易信號
- 情緒 Agent - 分析市場情緒並產生交易信號
- 基本面 Agent - 分析基本面數據並產生交易信號
- 技術分析 Agent - 分析技術指標（如 MACD、RSI、Bollinger Bands）等並產生交易信號
- 風險管理 Agent - 計算風險指標（如波動率、最大回撤等）並設定持倉限制
- 投資組合經理 - 作出最終交易決策並產生訂單
每個 Agent 都會展示其推理過程，讓你清楚了解它們的運作方式
This system employs several agents working together:
- Market Data Analyst - Gathers and preprocesses market data
- Valuation Agent - Calculates the intrinsic value of a stock and generates trading signals
- Sentiment Agent - Analyzes market sentiment and generates trading signals
- Fundamentals Agent - Analyzes fundamental data and generates trading signals
- Technical Analyst - Analyzes technical indicators and generates trading signals
- Risk Manager - Calculates risk metrics and sets position limits
- Portfolio Manager - Makes final trading decisions and generates orders
Clone the repository:
git clone https://github.com/Investment-Analyst/ai-hedge-fund.git
cd ai-hedge-fund- Install Poetry (if not already installed):
curl -sSL https://install.python-poetry.org | python3 -- Install dependencies:
poetry install- Set up your environment variables:
# Create .env file for your API keys
cp .env.example .env
export OPENAI_API_KEY='your-api-key-here' # Get a key from https://platform.openai.com/
export FINANCIAL_DATASETS_API_KEY='your-api-key-here' # Get a key from https://financialdatasets.ai/poetry run python src/main.py --ticker AAPLExample Output:
Final Result:
{
"action": "hold",
"quantity": 0,
"confidence": 0.63,
"agent_signals": [
{
"agent_name": "Valuation Analysis",
"signal": "bearish",
"confidence": 0.82
},
{
"agent_name": "Fundamental Analysis",
"signal": "bearish",
"confidence": 0.5
},
{
"agent_name": "Technical Analysis",
"signal": "bullish",
"confidence": 0.24
},
{
"agent_name": "Sentiment Analysis",
"signal": "bearish",
"confidence": 1.0
}
],
"reasoning": "The risk management constraints require a bearish action.
Valuation Analysis provides a strong bearish signal with 82% confidence due to significant overvaluation.
Fundamental Analysis also supports a bearish stance with concerns over growth and high price ratios.
Sentiment is unanimously bearish with 100% confidence.
Technical Analysis is bullish but holds the least weight and confidence.
Given these factors, the decision is to hold, with no buy or sell action possible due to portfolio constraints."
}
You can also specify a --show-reasoning flag to print the reasoning of each agent to the console.
poetry run python src/main.py --ticker AMD --show-reasoningYou can optionally specify the start and end dates to make decisions for a specific time period.
poetry run python src/main.py --ticker AAPL --start-date 2024-01-01 --end-date 2025-01-01 Example Output:
Final Result:
{
"action": "hold",
"quantity": 0,
"confidence": 0.82,
"agent_signals": [
{
"agent": "Technical Analysis",
"signal": "bullish",
"confidence": 0.24
},
{
"agent": "Fundamental Analysis",
"signal": "bearish",
"confidence": 0.50
},
{
"agent": "Sentiment Analysis",
"signal": "bearish",
"confidence": 1.00
},
{
"agent": "Valuation Analysis",
"signal": "bearish",
"confidence": 0.82
}
],
"reasoning": "The risk management constraints require a 'hold' action.
The dominant bearish signals from valuation (82% confidence) and sentiment (100% confidence) outweigh the bullish technical signal (24% confidence).
Fundamental analysis is also bearish. Thus, despite the technical bullish signal, the overall bearish sentiment and valuation drive the decision to hold."
poetry run python src/backtester.py --ticker AAPLExample Output:
Starting backtest...
Date Ticker Action Quantity Price Cash Stock Total Value
----------------------------------------------------------------------
2024-01-01 AAPL buy 519.0 192.53 76.93 519.0 100000.00
2024-01-02 AAPL hold 0 185.64 76.93 519.0 96424.09
2024-01-03 AAPL hold 0 184.25 76.93 519.0 95702.68
2024-01-04 AAPL hold 0 181.91 76.93 519.0 94488.22
2024-01-05 AAPL hold 0 181.18 76.93 519.0 94109.35
2024-01-08 AAPL sell 519 185.56 96382.57 0.0 96382.57
2024-01-09 AAPL buy 520.0 185.14 109.77 520.0 96382.57
You can optionally specify the start and end dates to backtest over a specific time period.
poetry run python src/backtester.py --ticker AAPL --start-date 2024-01-01 --end-date 2024-03-01ai-hedge-fund/
├── src/
│ ├── agents/ # Agent definitions and workflow
│ │ ├── fundamentals.py # Fundamental analysis agent
│ │ ├── market_data.py # Market data agent
│ │ ├── portfolio_manager.py # Portfolio management agent
│ │ ├── risk_manager.py # Risk management agent
│ │ ├── sentiment.py # Sentiment analysis agent
│ │ ├── state.py # Agent state
│ │ ├── technicals.py # Technical analysis agent
│ │ ├── valuation.py # Valuation analysis agent
│ ├── tools/ # Agent tools
│ │ ├── api.py # API tools
│ ├── backtester.py # Backtesting tools
│ ├── main.py # Main entry point
├── pyproject.toml
├── ...
When weighing the different signals for direction and timing:
1. Valuation Analysis (35% weight)
- Primary driver of fair value assessment
- Determines if price offers good entry/exit point
2. Fundamental Analysis (30% weight)
- Business quality and growth assessment
- Determines conviction in long-term potential
3. Technical Analysis (25% weight)
- Secondary confirmation
- Helps with entry/exit timing
4. Sentiment Analysis (10% weight)
- Final consideration
- Can influence sizing within risk limits
The decision process should be:
1. First check risk management constraints
2. Then evaluate valuation signal
3. Then evaluate fundamentals signal
4. Use technical analysis for timing
5. Consider sentiment for final adjustment
Provide the following in your output:
- "action": "buy" | "sell" | "hold",
- "quantity": <positive integer>
- "confidence": <float between 0 and 1>
- "agent_signals": <list of agent signals including agent name, signal (bullish | bearish | neutral), and their confidence>
- "reasoning": <concise explanation of the decision including how you weighted the signals>
Trading Rules:
- Never exceed risk management position limits
- Only buy if you have available cash
- Only sell if you have shares to sell
- Quantity must be ≤ current position for sells
- Quantity must be ≤ max_position_size from risk management"""