📈 Stock Analysis using MATLAB (ARIMA & LSTM)

This project applies time series forecasting techniques (ARIMA and LSTM models) to predict stock market prices using MATLAB.
It demonstrates how classical statistical methods and modern deep learning approaches compare in financial prediction tasks.

---

🚀 Features

• ARIMA Model

  • Captures linear trends & seasonality.
  • Useful for short-term forecasting.

• LSTM Model

  • Recurrent neural network specialized for sequential data.
  • Captures long-term dependencies in stock prices.
  • Outperforms ARIMA for highly volatile data.

• Comparison of Results

  • Forecast accuracy measured using RMSE & MAPE.
  • Graphical comparison of predicted vs. actual stock prices.

---

---

🔧 Requirements

• MATLAB R2021a or later (older versions may work).
• Toolboxes:
  • Econometrics Toolbox
  • Deep Learning Toolbox

---

📊 Dataset

You can download historical stock price data from the NASDAQ website:
👉 NASDAQ Historical Data

Example:

1. Search for a ticker symbol (e.g., AAPL for Apple, TSLA for Tesla).
2. Go to the “Historical Data” tab.
3. Select the time period (e.g., 5 years).
4. Click Download (CSV file).
5. Place the CSV file in the data/ folder.

---

▶️ How to Run

1. Clone or download this repository.
2. Download your dataset from NASDAQ and save it in /data.
3. Open codefile.m in MATLAB.
4. Run the live script step by step to:
   • Load dataset
   • Train ARIMA & LSTM models
   • Compare results

---

📊 Sample Results

Placeholder for graphs — upload your PNG plots here, e.g.:

• ARIMA forecast vs. actual
• LSTM forecast vs. actual

---

🔮 Future Improvements

• Hybrid ARIMA-LSTM model.
• Transformer-based time series forecasting.
• Multi-stock portfolio forecasting.

---

✨ Author

Project co-developed as part of a Joint academic project.