Stock Prediction ML Model

A machine learning application for predicting stock prices using LSTM (Long Short-Term Memory) neural networks. This project combines historical stock data analysis with deep learning to forecast future price trends.

🎯 Features

• Real-time Stock Data: Fetches current and historical stock data using yfinance API
• LSTM Neural Network: Advanced deep learning model for time series forecasting
• Interactive Visualizations:
  • Heikin-Ashi candlestick charts with volume indicators
  • Moving average analysis (100-day and 200-day MA)
  • Predictions vs Original price comparison
• Pre-trained Models: Uses saved Keras models for fast predictions
• User-friendly Interface: Built with Streamlit for easy interaction

📊 Project Structure

Stock Prediction ML Model/
├── README.md                          # This file
├── requirements.txt                   # Python dependencies
├── LICENSE                            # Project license
├── src/
│   ├── __init__.py
│   └── Stockinfo.py                   # Main Streamlit application
├── models/
│   └── keras_model.h5                 # Pre-trained LSTM model
├── notebooks/
│   ├── LSTM model.ipynb               # Model training notebook
│   ├── LSTM model-checkpoint.ipynb    # Checkpoint notebook
│   ├── keras_model.h5                 # Saved model
│   └── my_model_keras.h5              # Alternative model
├── docs/
│   ├── architecture.md                # Detailed architecture documentation
│   └── README.md                      # Additional documentation
├── assets/                            # Project assets
└── price_pred_venv_fixed/             # Virtual environment (optional)

🚀 Quick Start

Prerequisites

• Python 3.8 or higher
• pip or conda package manager

Installation

1. Clone or download the project

   cd "Stock Prediction ML Model"

2. Create a virtual environment (optional but recommended)

   python -m venv venv
   source venv/Scripts/activate  # On Windows
   # or
   source venv/bin/activate      # On macOS/Linux

3. Install dependencies

   pip install -r requirements.txt

Running the Application

streamlit run src/Stockinfo.py

The app will open in your default browser at http://localhost:8501

📈 How It Works

Data Collection

• Fetches historical OHLCV (Open, High, Low, Close, Volume) data from Yahoo Finance
• Allows custom date ranges for analysis

Data Processing

• Normalization: MinMax scaling (0-1 range) for stable neural network training
• Heikin-Ashi Candles: Smoothed candlestick representation for better trend visibility
• Sequence Creation: Converts time series into overlapping windows for LSTM training

Model Architecture

Input Layer (100 timesteps × 1 feature)
    ↓
LSTM Layer (50 units) + Dropout (0.2)
    ↓
LSTM Layer (60 units) + Dropout (0.3)
    ↓
LSTM Layer (80 units) + Dropout (0.4)
    ↓
LSTM Layer (120 units) + Dropout (0.5)
    ↓
Dense Layer (1 unit)
    ↓
Output (Predicted Price)

Predictions

• Uses 70% of data for training, 30% for testing
• Predicts next day stock prices based on 100-day historical patterns
• Compares predictions against actual prices

💻 Dependencies

Package	Purpose
numpy	Numerical computing
pandas	Data manipulation and analysis
matplotlib	Static visualizations
seaborn	Statistical data visualization
yfinance	Fetch stock data from Yahoo Finance
tensorflow	Deep learning framework with Keras
streamlit	Interactive web app framework
scikit-learn	Data preprocessing (MinMaxScaler)
altair	Interactive charting library

📝 Usage Guide

Using the Application

1. Enter Stock Ticker: Input any valid stock symbol (e.g., AAPL, GOOGL, MSFT)
2. Select Date Range: Choose start and end dates for historical analysis
3. View Analysis:
   • Data statistics and descriptions
   • Closing price trends
   • Heikin-Ashi candlestick charts
   • Moving averages (100-day, 200-day)
   • Prediction results

Example Tickers

• AAPL (Apple)
• GOOGL (Google)
• MSFT (Microsoft)
• TSLA (Tesla)
• AMZN (Amazon)

🔧 Training a New Model

To train a new LSTM model with different parameters:

1. Open notebooks/LSTM model.ipynb
2. Modify the model architecture or hyperparameters
3. Run the notebook cells
4. Save the trained model to models/keras_model.h5

⚠️ Important Notes

• Internet Required: The app needs an active internet connection to fetch stock data
• Model Performance: Past performance doesn't guarantee future results
• Data Limitations: 'Adj Close' may not be available for all tickers; the app uses 'Close' as fallback
• Prediction Accuracy: LSTM models work best with sufficient historical data (ideally 5+ years)

📊 Visualizations

1. Data Statistics

• Summary statistics of OHLCV data
• Data availability and quality checks

2. Heikin-Ashi Candlesticks

• Green candles: Bullish (closing > opening)
• Red candles: Bearish (closing < opening)
• Volume bars synchronized with price movements

3. Moving Averages

• 100-day moving average for short-term trends
• 200-day moving average for long-term trends
• Helps identify trend direction and support/resistance

4. Predictions

• Comparison graph of actual vs predicted prices
• Performance metrics visualization

🎓 Learning Resources

• LSTM Networks: Understanding LSTMs
• Keras/TensorFlow: TensorFlow Documentation
• Time Series Forecasting: Forecasting Principles

🤝 Contributing

Feel free to fork, modify, and improve this project. Some ideas for enhancement:

• Add more technical indicators (RSI, MACD, Bollinger Bands)
• Implement ensemble models
• Add sentiment analysis from news/social media
• Create model comparison functionality
• Add portfolio prediction for multiple stocks

📄 License

This project is licensed under the terms specified in the LICENSE file.

🐛 Troubleshooting

TensorFlow Import Error

pip install tensorflow --upgrade

yfinance Data Issues

• Check internet connection
• Verify ticker symbol is valid
• Try a different date range

Streamlit Not Found

pip install streamlit

Model Loading Error

• Ensure models/keras_model.h5 exists
• If missing, the app will train a new model automatically

📞 Support

For issues or questions:

1. Check the docs/architecture.md for technical details
2. Review the notebooks in the notebooks/ directory
3. Examine the Streamlit app logs in the terminal

✨ Future Enhancements

• Multi-stock portfolio analysis
• Real-time predictions
• Risk assessment metrics
• Alternative models (GRU, Transformer, Prophet)
• Model performance metrics (RMSE, MAE)
• Data caching for faster reloads
• Advanced charting with technical indicators

---

Last Updated: December 2025