FNEPSE - Stock Market Analyzer & Predictor

A comprehensive AI-powered stock market analysis and trading signal prediction system that combines CNN and LSTM neural networks with technical analysis. The system includes real-time NEPSE (Nepal Stock Exchange) data scraping, pattern recognition, and both command-line and web interfaces.

🚀 Features

Core Functionality

• AI-Powered Predictions: CNN-LSTM hybrid model for trading signal generation
• Pattern Recognition: Candlestick pattern analysis using deep learning
• Technical Analysis: 15+ technical indicators (RSI, MACD, Moving Averages, etc.)
• Real-time Data: Live NEPSE market data scraping and analysis
• Multi-timeframe Analysis: Support for various timeframes and lookback periods

Interfaces

• Web Application: Modern, responsive web interface with Bootstrap 5
• CLI Tool: Command-line interface for batch processing and automation
• REST API: RESTful endpoints for integration with other systems

Data Sources

• NEPSE Integration: Real-time data from Nepal Stock Exchange
• Kaggle Datasets: Historical US stock and ETF data for training
• Custom Data Input: Support for manual OHLCV data entry

📊 Model Architecture

The system uses a sophisticated hybrid architecture:

1. CNN Layer: Extracts local patterns from candlestick formations
2. LSTM Layer: Captures temporal dependencies and market trends
3. Technical Indicators: Enhanced feature engineering with 10+ indicators
4. Ensemble Methods: Combines multiple prediction strategies

🛠️ Installation

Prerequisites

• Python 3.8 or higher
• pip package manager
• 4GB+ RAM (recommended for model training)

Quick Setup

1. Clone the repository:

   git clone https://github.com/sthaarwin/FNEPSE.git
   cd FNEPSE

2. Install dependencies:

   pip install -r requirements.txt

3. Download and preprocess data (optional - for training):

   python src/data_loader.py
   python src/preprocess.py

🚀 Usage

Web Application

Start the web interface:

python app.py

Then open your browser to http://localhost:5000

Enhanced Web App with NEPSE Integration

For the full-featured web app with real-time NEPSE data:

python web_app.py

Access at http://localhost:5001

Command Line Interface

For direct predictions:

python trading_signal_predictor.py

API Usage

Make predictions via REST API:

curl -X POST http://localhost:5000/predict \
  -H "Content-Type: application/json" \
  -d '{
    "ohlcv_data": [
      {"open": 100, "high": 105, "low": 98, "close": 103},
      {"open": 103, "high": 107, "low": 101, "close": 106}
    ]
  }'

📁 Project Structure

FNEPSE/
├── app.py                         # Main Flask web application
├── web_app.py                     # Enhanced web app with NEPSE integration
├── trading_signal_predictor.py    # Core prediction engine
├── config.yaml                    # Configuration settings
├── requirements.txt               # Python dependencies
├── data/                          # Data storage
│   ├── raw/                       # Raw market data
│   ├── processed/                 # Cleaned datasets
│   └── preprocessed/              # Model-ready data
├── models/                        # Trained models
│   ├── candlestick_cnn_lstm.h5    # Pattern recognition model
│   ├── trading_signals_model.h5   # Trading signal model
│   └── scaler.pkl                 # Feature scaler
├── scraper/                       # NEPSE data scraper
│   └── nepse_scraper/             # Web scraping utilities
├── src/                           # Source code
│   ├── data_loader.py             # Data downloading and loading
│   └── preprocess.py              # Data preprocessing pipeline
├── notebooks/                     # Jupyter notebooks
├── templates/                     # HTML templates
├── static/                        # Static web assets
└── utils/                         # Utility functions

🎯 Trading Signals

The system generates three types of signals:

• BUY (2): Strong bullish indicators suggest potential upward movement
• HOLD (1): Mixed or neutral signals, maintain current position
• SELL (0): Bearish indicators suggest potential downward movement

Each prediction includes:

• Signal strength (0-2)
• Confidence score (0-1)
• Detailed explanation of factors
• Technical indicator analysis

📈 Technical Indicators

The system analyzes multiple technical indicators:

• Trend: Moving Averages (5, 20), EMA (12, 26)
• Momentum: RSI, MACD, Stochastic Oscillator
• Volatility: Standard Deviation, Bollinger Bands
• Volume: Volume surge analysis
• Price Action: Support/Resistance levels, Price spreads

🔧 Configuration

Customize the system behavior via config.yaml:

model:
  lookback_period: 20
  confidence_threshold: 0.7
  
trading:
  risk_tolerance: "medium"
  max_positions: 10
  
data:
  update_frequency: "daily"
  data_sources: ["nepse", "kaggle"]

🚦 API Endpoints

Endpoint	Method	Description
/	GET	Web interface homepage
/predict	POST	Generate trading signal
/market_data	GET	Current NEPSE market data
/predict_for_ticker	POST	Predict for specific stock
/recommendations	GET	AI-generated stock recommendations
/health	GET	System health check

🔒 Risk Warning

⚠️ Important Disclaimer: This tool is for educational and research purposes only.

• Past performance does not guarantee future results
• Always conduct your own research before making investment decisions
• Consider consulting with financial advisors for investment advice
• The developers are not responsible for any financial losses

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Nepal Stock Exchange (NEPSE) for market data
• Kaggle community for stock datasets
• TensorFlow and Keras teams for ML frameworks
• Flask community for web framework

📧 Contact

• Author: Arwin Shrestha
• GitHub: @sthaarwin
• Project Link: https://github.com/sthaarwin/FNEPSE

---

Made with ❤️ for the trading community