This repository contains machine learning models trained on NSE (National Stock Exchange of India) bhavcopy data for stock price prediction. The models are trained on data from January 1, 2025 to August 20, 2025, covering 3,257 unique stock symbols across 158 trading days.

Model Information

Data Source

Source: NSE Bhavcopy (daily market data)
Period: January 1, 2025 - August 20, 2025
Coverage: 3,257 Indian equity stocks
Features: 16 technical and price-based features

Features Used

Price Data: Open, High, Low, Close, Last Price, Previous Close
Technical Indicators: Price Range, Price Change, Price Change %, Volume-Price Ratio
Moving Averages: SMA_5, SMA_20
Momentum: Price Momentum, Volume MA, Volume Ratio
Volume: Total Trading Volume, Total Trading Value

Models Available

RandomForest - Ensemble of 100 decision trees
GradientBoosting - Gradient boosting with 100 estimators
LinearRegression - Linear regression model
Ridge - Ridge regression with L2 regularization
Lasso - Lasso regression with L1 regularization
SVR - Support Vector Regression
XGBoost - Extreme Gradient Boosting
LightGBM - Light Gradient Boosting Machine

Usage

Installation

pip install -r requirements.txt

Quick Start

python

from inference import StoxChaiStockPredictor
import numpy as np

# Initialize predictor
predictor = StoxChaiStockPredictor()

# Prepare your data (16 features in the same order as training)
features = [100.0, 105.0, 98.0, 102.0, 100.0, 7.0, 2.0, 2.0, 1.5, 
           101.0, 100.5, 0.01, 1000.0, 1.2, 1200.0, 120000.0]

# Single model prediction
prediction = predictor.predict(features, "randomforest")
print(f"Predicted stock price: ₹{prediction:.2f}")

# All models prediction
all_predictions = predictor.predict_all_models(features)
print(f"Ensemble prediction: ₹{all_predictions['ensemble']:.2f}")

Input Format

Your input should be a 1D array with exactly 16 features in this order:

OpnPric (Open Price)
HghPric (High Price)
LwPric (Low Price)
LastPric (Last Price)
PrvsClsgPric (Previous Close Price)
Price_Range
Price_Change
Price_Change_Pct
Volume_Price_Ratio
SMA_5
SMA_20
Price_Momentum
Volume_MA
Volume_Ratio
TtlTradgVol (Total Trading Volume)
TtlTrfVal (Total Trading Value)

Performance Metrics

Training Data: 464,548 samples
Validation: 20% holdout test set
Models trained successfully: 8/8
All models converged without errors

License

MIT License - See LICENSE file for details

Citation

If you use these models in your research, please cite:

@software{stoxchai_nse_predictor,
  title={StoxChai NSE Stock Price Prediction Models},
  author={StoxChai Team},
  year={2025},
  url={https://huggingface.co/thoutam/stoxchai-nse-predictor}
}

Support

For questions or issues, please open an issue on this repository.