About Qunex Trade

What is Qunex Trade?

Qunex Trade is an advanced machine learning system designed to predict penny stock surges (50%+ intraday gains) with exceptional accuracy. Using ensemble learning and 115+ engineered features, the model achieves a 0.9789 ROC-AUC score and maintains a 70%+ success rate in real-world trading scenarios.

How It Works

1. Data Collection & Feature Engineering

The system continuously monitors 522 penny stocks, collecting historical price, volume, and technical indicator data. Advanced feature engineering creates 115+ predictive features including volatility patterns, momentum indicators, and price-volume relationships.

2. Machine Learning Ensemble

Four state-of-the-art ML models work together:

• XGBoost Advanced - Gradient boosting with custom parameters (ROC-AUC: 0.9783)
• LightGBM Advanced - Fast gradient boosting (ROC-AUC: 0.9699)
• RandomForest Advanced - Ensemble decision trees (ROC-AUC: 0.9740)
• GradientBoosting - Classic boosting (ROC-AUC: 0.9749)

The models are weighted and combined to achieve a final ensemble ROC-AUC of 0.9789.

3. Signal Generation (4:05 PM Daily)

Every day at 4:05 PM ET (after market close), the system automatically:

• Downloads the latest market data for all 522 stocks
• Calculates all 115+ features for each stock
• Runs predictions through the ensemble model
• Filters for ultra-high confidence signals (90%+ probability)
• Saves signals for the next trading day

4. Performance Tracking

The system tracks every signal's actual performance by comparing predicted vs. actual results. Success is measured by 50%+ intraday gains (buy at open, sell at close). All signals remain active over weekends and holidays until the next trading day.

Key Features

Technology Stack

Machine Learning

Data & APIs

Web Application

Automation

Disclaimer