🎯 Predict T20I Cricket Match Winners with 95% Accuracy using Random Forest!
🚀 AI-powered cricket match prediction system that analyzes historical T20I data to predict match outcomes with exceptional accuracy.
This machine learning project uses advanced algorithms to forecast T20 International cricket match winners based on:
- Historical match statistics
- Venue performance data
- Toss decisions impact
- Team-specific analytics
AI Prediction 95% Accuracy Rate |
Interactive GUI User-Friendly Interface |
Venue Analysis Ground-Specific Stats |
Match History Historical Data Insights |
| Model | Accuracy | Status |
|---|---|---|
| 🌟 Random Forest | 95.00% | ✅ Best Performer |
| ⚡ XGBoost | 66.94% | |
| 📈 SVC | 11.00% | ❌ Poor |
| 📊 Logistic Regression | 10.78% | ❌ Poor |
| 🔄 AdaBoost | 9.21% | ❌ Poor |
✅ Handles categorical variables excellently
✅ Captures non-linear relationships
✅ Prevents overfitting with ensemble approach
✅ Perfect for multi-class classification
| Feature | Description | Type |
|---|---|---|
| Bat First | Toss winner team | Categorical |
| Bat Second | Team batting second | Categorical |
| Venue | Match ground/stadium | Categorical |
| Winner | Match winner (target) | Label |
🎯 Input: Team 1, Team 2, Venue, Toss Winner
📤 Output: Predicted match winner
📊 View: Past match statistics for selected teams
✅ Verify: Model accuracy against actual results
📊 Analyze: Ground-specific team performance
📈 Trends: Historical venue statistics
graph LR
A[📊 Raw Data] --> B[🔄 Preprocessing]
B --> C[⚙️ Feature Engineering]
C --> D[🎯 Model Training]
D --> E[🏆 Random Forest]
E --> F[💻 GUI Interface]
- 📊 Data Preprocessing - Ball-by-ball to match-level aggregation
- 🏷️ Feature Engineering - One-hot encoding & label encoding
- 📏 Scaling - StandardScaler implementation
- 🎯 Training - 80/20 train-test split
- 💻 GUI Development - Interactive prediction interface
📅 Test Case: India vs Pakistan at Dubai International Stadium
📊 Historical Data: Pakistan won both previous matches when India batted first
🎯 Model Prediction: Pakistan (Winner)
✅ Validation: Matches historical trend!
| Metric | Value |
|---|---|
| 🎯 Accuracy | 95% |
| 📊 Models Tested | 5 |
| 💻 GUI Features | 3 |
| 📅 Venues Covered | Multiple International Grounds |
# Clone the repository
git clone <https://github.com/MudasirNaeem1/MachineLearning-T20I-Matches-Prediction.git>
# Install dependencies
pip install -r requirements.txt
# Run the application
ML_PROJECT (T20I CRICKET PREDICTION MODEL).ipynbNATIONAL UNIVERSITY OF COMPUTER & EMERGING SCIENCES
📍 Karachi Campus | 🎓 BAI-5A
👤 Instructor: Sir Usama Bin Umar
📅 Date: December 14, 2024
Interested in educational AI research? ⭐ Star this repository!
Have questions about the implementation? 💭 Let's discuss!
Found this project interesting? ⭐ Star the repository!
Have suggestions? 💭 Reach out!