Crash Predictor ML — OCR + Machine Learning Experiment

This project is an experimental machine learning bot that:

• Uses OCR (EasyOCR + OpenCV) to read crash multipliers live from the screen
• Tracks game history in real time
• Trains a RandomForest Classifier + Regressor (Scikit-Learn)
• Predicts the probability of high multipliers
• Provides a Tkinter-based GUI for live monitoring, settings, and training
• Supports auto-training and CSV logging for long sessions

⚠️ Disclaimer:
This project is for educational purposes only.
It is not a guaranteed profit tool and should not be used for gambling.
The purpose is to demonstrate OCR, real-time data capture, and ML prediction pipelines.

---

🚀 Features

• OCR Integration: EasyOCR + OpenCV for screen capture and text recognition
• Machine Learning: RandomForest Classifier & Regressor with auto-training
• GUI Dashboard: Tkinter interface with live logs and model management
• Customizable Settings: History window, feature rounds, training intervals
• Data Logging: Saves predictions, outcomes, and retraining history to CSV

---

📦 Requirements

• Python 3.10+
• Packages: numpy, pandas, scikit-learn, easyocr, opencv-python, tkinter, joblib

Install with:

pip install -r requirements.txt