📈 Stock Market Prediction — LSTM

A professional-grade, interactive stock price prediction web app built with LSTM (Long Short-Term Memory) neural networks and deployed via Streamlit. Built this project as my final project for the semester.

---

Demo

Features

Feature	Details
Interactive Technical Chart	Candlestick chart with MA20, MA50, MA200, Bollinger Bands, RSI, MACD
LSTM Model	2-layer LSTM with Dropout, Early Stopping, and train/test split
Model Metrics	RMSE, MAE, MAPE, and Direction Accuracy on the test set
Future Forecasting	Predict up to 90 business days into the future
Configurable Hyperparameters	Look-back window, LSTM units, dropout, epochs, batch size
Company Info	Real-time KPIs: latest close, 52-week high/low, market cap, sector
Raw Data Table	View last 50 rows of enriched DataFrame
Dark Theme UI	Professional Plotly dark theme throughout

---

Demo

Ticker:     AAPL / MSFT / TSLA / NVDA / any Yahoo Finance symbol
Date Range: Configurable (default 2018–2025)
Forecast:   Up to 90 days ahead

---

Installation

1. Clone the repository

git clone https://github.com/<your-username>/stock-market-prediction.git
cd stock-market-prediction

2. Create a virtual environment (recommended)

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

3. Install dependencies

pip install -r requirements.txt

4. Run the app

streamlit run sm.py

The app opens automatically at http://localhost:8501.

---

📁 Project Structure

stock-market-prediction/
│
├── sm.py               # Main Streamlit application
├── requirements.txt    # Python dependencies
└── README.md           # This file

---

Model Architecture

Input → LSTM(units, return_sequences=True)
      → Dropout
      → LSTM(units/2)
      → Dropout
      → Dense(25, relu)
      → Dense(1)

Loss:      Mean Squared Error
Optimizer: Adam
Callbacks: EarlyStopping (patience=8, restore_best_weights=True)

---

Technical Indicators

Indicator	Period
Simple Moving Average	MA20, MA50, MA200
Bollinger Bands	20-day, ±2σ
RSI	14-day
MACD	EMA(12) − EMA(26) + Signal(9)

---

Hyperparameters (Sidebar)

Parameter	Default	Range
Look-back Window	60 days	30–120
LSTM Units	100	32–256
Dropout Rate	0.2	0.0–0.5
Max Epochs	50	10–100
Batch Size	64	16–128
Train Split	80%	60–90%
Future Forecast	30 days	0–90

---

Disclaimer

This project is for educational purposes only. Stock market predictions are inherently uncertain and should not be used as financial advice. Always consult a licensed financial advisor before making investment decisions.

---

📄 License

This project is licensed under the MIT License.

---

Acknowledgements

• yfinance — Yahoo Finance market data
• TensorFlow / Keras — Deep learning framework
• Streamlit — App framework
• Plotly — Interactive charts