Stock Analysis Prediction Model

🎯 Project Overview

This project implements a stock price prediction model using two different machine learning approaches: linear regression and Long-Short-Term Memory (LSTM) neural networks. The goal is to provide predictive insights into stock price movements using historical data from Yahoo Finance.

Research Paper - https://1drv.ms/b/c/87e0048ba6376a7d/EUPKbpK_j6dIgtuLYuGBNnQBpwREqJubv3_DSX9T34SmkQ Video Demonstration - https://youtu.be/z8sXhWrwU0o

💻 Features

• Automated stock data retrieval using yfinance
• Two prediction models:
  • Linear Regression
  • LSTM (Long Short-Term Memory) Neural Network
• Comprehensive data preprocessing
• Model training and evaluation
• Comparative analysis of prediction performance

📚 Prerequisites

• Python
• Tensorflow
• PyTorch
• yfinance
• pandas
• numpy
• Matlpotlib
• joblib
• sklearn

🚀 Installation

1. Clone the repository:

git clone https://github.com/yourusername/Stock_Analysis_Prediction_Model.git
cd Stock_Analysis_Prediction_Model

2. Create a virtual environment (optional but recommended):

python -m venv venv
source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

3. Install required packages:

pip install -r requirements.txt

📂 Project Structure

Stock_Analysis_Prediction_Model/
│
├── data/                    # Raw and processed stock data
├── src/                     # Source code for data fetching and model training
├── models/                  # Saved trained models
├── tests/                   # Unit tests for various components
├── images/                # Model performance visualization
├── requirements.txt         # Project dependencies
└── main.py                  # Entry point for running the prediction model

🔧 Usage

Fetching Stock Data

To fetch historical stock data:

python src/fetch_data.py

Training Models & Running Predictions

To train both Linear Regression and LSTM models & run predictions and compare model performance:

python src/main.py

⚖️ Model Comparison

The project compares two machine learning models:

1. Linear Regression

   • Simple, interpretable model
   • Works well with linear relationships
   • Faster training time

2. LSTM Neural Network

   • Captures complex temporal dependencies
   • Better at handling sequential data
   • More complex architecture

📏 Performance Metrics

The performance of each model is evaluated using:

• Mean Squared Error (MSE)
• Root Mean Squared Error (RMSE)
• Mean Absolute Error (MAE)
• R-squared (R²) Score

Detailed performance metrics are visualized in the pictures/ directory.

🧪 Testing

Run the test suite to verify model and data processing functionality:

python -m pytest tests/

🤝 Contributing

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

⚠️ Limitations

• Stock predictions are inherently probabilistic
• Model performance depends on market conditions
• Past performance does not guarantee future results

📝 License

Distributed under the MIT License. See LICENSE for more information.

📫 Contact

Harshit Kushwaha 🧑‍💻
Developer

📧 [email protected]

---