AI-powered stock prediction engine using machine learning and technical analysis.

• Real-time Stock Predictions - Predict price movements with confidence scores
• Technical Analysis - RSI, MACD, Bollinger Bands, and more
• AI Insights - Powered by Google Gemini for intelligent market commentary
• Batch Processing - Analyze entire portfolios at once
• Historical Tracking - Track prediction accuracy over time

• Python 3.8+ (you have 3.13.9 ✓)
• pip (package installer)
• Google Gemini API key (already configured ✓)

cd C:\Users\2cody\source\repos
git clone <your-ml-service-repo>  # Or create new folder
cd Nexus-Signal-ML

# Run setup script (Windows)
setup.bat

The setup script will:

• Create a Python virtual environment
• Install all dependencies (TensorFlow, Flask, etc.)
• Create necessary directories

# Activate virtual environment
venv\Scripts\activate.bat

# Start the Flask server
python app.py

The service will start on http://localhost:5001

Open your browser and go to:

http://localhost:5001/health

You should see:

{
  "status": "healthy",
  "service": "ML Prediction Service",
  "version": "1.0.0"
}

GET /health

POST /predict
Content-Type: application/json

{
  "symbol": "AAPL",
  "days": 7
}

Response:

{
  "symbol": "AAPL",
  "current_price": 178.32,
  "prediction": {
    "direction": "UP",
    "confidence": 72.5,
    "target_price": 182.15,
    "price_change_percent": 2.15,
    "timeframe_days": 7
  },
  "signals": [
    "RSI oversold (bullish)",
    "MACD bullish crossover"
  ],
  "technical_analysis": {
    "rsi": 42.3,
    "macd_signal": "Bullish",
    "volatility": 28.5,
    "volume_status": "High"
  }
}

POST /predict/batch
Content-Type: application/json

{
  "symbols": ["AAPL", "GOOGL", "MSFT"],
  "days": 7
}

POST /analyze
Content-Type: application/json

{
  "symbol": "AAPL"
}

Response includes:

• Prediction
• AI-generated insights from Google Gemini
• Detailed technical analysis
• Trading recommendations

Edit .env file:

PORT=5001
GOOGLE_API_KEY=your_api_key_here
MODEL_PATH=./models/saved_models
LOG_LEVEL=INFO

• Uses yfinance to get real-time stock data
• Fetches 6 months of historical data by default

• Calculates 10+ technical indicators
• RSI, MACD, Bollinger Bands, ATR, OBV, Stochastic

• Rule-based predictions using technical indicators
• Confidence scoring system
• Can be upgraded to LSTM/TensorFlow models

• Google Gemini generates human-readable insights
• Explains "why" behind predictions
• Provides actionable recommendations

The prediction engine uses a weighted scoring system:

Score = RSI_signal + MACD_signal + BB_signal + Momentum_signal + Volume_signal

If score >= 2:  Direction = UP,   Confidence = 60-85%
If score <= -2: Direction = DOWN, Confidence = 60-85%
Else:           Direction = NEUTRAL, Confidence = ~50%

• RSI < 30: Oversold (Bullish) +2 points
• RSI > 70: Overbought (Bearish) -2 points
• MACD > Signal: Bullish +1 point
• Price near lower BB: Bullish +1 point
• High volume: Momentum +0.5 points

Add to your Nexus-Signal-Server:

// server/routes/predictionsRoutes.js
const express = require('express');
const router = express.Router();
const axios = require('axios');
const auth = require('../middleware/authMiddleware');

const ML_SERVICE_URL = 'http://localhost:5001';

router.post('/predict', auth, async (req, res) => {
    try {
        const { symbol } = req.body;
        const response = await axios.post(`${ML_SERVICE_URL}/predict`, { symbol });
        res.json(response.data);
    } catch (error) {
        res.status(500).json({ error: 'Prediction service unavailable' });
    }
});

module.exports = router;

1. Upgrade to Deep Learning

   • Implement LSTM neural networks
   • Train on larger datasets
   • Add sentiment analysis from news

2. Add More Features

   • Options pricing models
   • Portfolio optimization
   • Risk analysis

3. Improve Accuracy

   • Collect prediction results
   • Calculate actual accuracy
   • Retrain models based on performance

ml-service/
├── app.py                      # Flask API server
├── requirements.txt            # Python dependencies
├── .env                        # Configuration
├── models/
│   ├── predictor.py           # ML prediction engine
│   └── saved_models/          # Trained models
├── utils/
│   ├── technical_indicators.py # Technical analysis
│   ├── market_data.py         # Data fetching
│   └── ai_insights.py         # Gemini AI integration
└── logs/                      # Service logs

Edit utils/technical_indicators.py:

def calculate_your_indicator(self, data):
    # Your calculation here
    return indicator_values

# models/trainer.py (create this file)
from predictor import StockPredictor

predictor = StockPredictor()
symbols = ['AAPL', 'GOOGL', 'MSFT', 'AMZN', 'TSLA']
predictor.train_model(symbols)

venv\Scripts\activate.bat
pip install -r requirements.txt

Change port in .env:

PORT=5002

1. Get key from: https://aistudio.google.com/apikey
2. Update .env file
3. Restart service

• TensorFlow Documentation
• yfinance Documentation
• Google Gemini API
• Technical Analysis Library

Need help? Check the logs:

tail -f logs/ml_service.log

MIT License - Built for Nexus Signal

Ready to predict the future? Let's go! 🚀