Professional-grade quantitative trading analytics powered by machine learning
INVESTO is an advanced stock prediction system that combines:
- Time Series Forecasting (ARIMA, SARIMAX)
- Ensemble Machine Learning (Gradient Boosting, XGBoost, LightGBM)
- Technical Analysis Indicators (32+ features)
- Interactive Visualization
Prototype Built for traders, analysts, and quantitative researchers to make data-driven investment decisions.
| Feature | Description | Technology |
|---|---|---|
| Multi-Model Forecasting | Ensemble of 5 prediction models with consensus analysis | ARIMA, SARIMAX, GB, XGB, LGB |
| Technical Analysis | 50+ indicators including RSI, MACD, Bollinger Bands | TA-Lib, Pandas |
| Professional Dashboard | Institutional-grade visualization | Plotly, Streamlit |
| Automated Feature Engineering | Dynamic feature generation pipeline | Scikit-learn |
| Model Comparison | Side-by-side performance metrics | MLflow, Pickle |
Real-time candlestick charts with overlay indicators
RMSE, MAE, and R² metrics across all models
Ensemble forecast with confidence intervals
- Clone the repository:
git clone https://github.com/jayash1973/INVESTO-Stock-Predictor.git
cd INVESTO-Stock-Predictor- Create virtual environment:
python -m venv venv
source venv/bin/activate # Linux/Mac
venv\Scripts\activate # Windows- Install dependencies:
pip install -r requirements.txtINVESTO-Stock-Predictor/
├── artifacts/ # Serialized models and processed data
│ ├── train_test_splits/ # Time-based splits
│ ├── engineered_data/ # Feature-enhanced datasets
│ └── scalers/ # Normalization objects
├── logs/ # Training logs
├── src/ # Core application code
│ ├── components/ # Modular ML components
│ ├── pipeline/ # Training/prediction workflows
│ └── utils.py # Helper functions
├── app.py # Streamlit application
└── requirements.txt # Dependency specification
graph TD
A[Raw OHLCV Data] --> B{Feature Engineering}
B --> C[Technical Indicators]
B --> D[Statistical Features]
B --> E[Temporal Features]
C & D & E --> F[Model Training]
F --> G[ARIMA]
F --> H[SARIMAX]
F --> I[Gradient Boosting]
F --> J[XGBoost]
F --> K[LightGBM]
G & H & I & J & K --> L[Ensemble Prediction]
L --> M[Dashboard Visualization]
- Run the Streamlit app:
streamlit run app.py-
Configure Analysis:
- Select stock ticker (AAPL, GOOGL, etc.)
- Choose date range (1Y, 5Y, custom)
- Toggle technical indicators
-
Generate Predictions:
- Click "Generate Predictions" button
- View model consensus forecast
- Analyze individual model performance
| Category | Indicators | Description |
|---|---|---|
| Trend | SMA(50,200), EMA(20) | Moving averages for trend identification |
| Momentum | RSI(14), MACD(12,26,9) | Speed and direction of price changes |
| Volatility | Bollinger Bands, ATR | Price fluctuation measurements |
| Volume | OBV, VWAP | Trading activity analysis |
| Cycle | Fourier Transforms | Periodic pattern detection |
| Model | Parameters | Use Case |
|---|---|---|
| ARIMA | (p=2,d=1,q=2) | Baseline price prediction |
| SARIMAX | (2,1,2)x(1,1,1,5) | Seasonal pattern capture |
| Model | Key Features | Strengths |
|---|---|---|
| Gradient Boosting | n_estimators=200, max_depth=4 | Handles non-linear relationships |
| XGBoost | learning_rate=0.05, subsample=0.8 | Regularization prevents overfit |
| LightGBM | num_leaves=31, min_data_in_leaf=5 | Fast training on large datasets |
Top predictive features across models
This project is licensed under the MIT License - see the LICENSE file for details.
We welcome contributions! Please follow these steps:
- Fork the repository
- Create your feature branch (
git checkout -b feature/AmazingFeature) - Commit your changes (
git commit -m 'Add some AmazingFeature') - Push to the branch (
git push origin feature/AmazingFeature) - Open a Pull Request
For professional inquiries: