🏏 IPL Match Prediction System

An AI-powered system that predicts IPL cricket match outcomes with 73.85% accuracy using machine learning and historical data analysis.

🚀 Overview

This project analyzes 16+ years of IPL data (2008-2024) to predict match winners using multiple ML algorithms. The system considers team performance, venue statistics, head-to-head records, recent form, and toss decisions to make intelligent predictions.

✨ Key Features

• High Accuracy Models: Neural Networks (73.85%), Gradient Boosting (72.94%)
• Interactive CLI: User-friendly prediction interface
• Smart Analytics: Venue advantage analysis and team form tracking
• Multiple Algorithms: Random Forest, XGBoost, Neural Networks, Gradient Boosting
• Real-time Predictions: Input teams, venue, and toss details for instant results

🏗️ Project Structure

📁 ipl-prediction/
├── 📊 data/
│   ├── raw/           # Original CSV datasets
│   └── processed/     # Cleaned feature data
├── 🤖 models/         # Trained ML models (.pkl files)
├── 📈 results/        # Performance metrics & visualizations
├── 🎯 main.py         # Complete pipeline runner
├── 🔧 data_integration.py    # Data preprocessing
├── 🏋️ model_training.py      # ML model training
├── 🧠 neural_network_sklearn.py  # Neural network training
├── 💬 predict_cli.py         # Interactive prediction tool
└── ⚡ predict_match.py       # Core prediction engine

🎯 Quick Start

1. Setup Data Files

Place these CSV files in data/raw/:

• cricket_data_2025.csv (player statistics)
• ipl_matches_detail_2008_2024.csv (match details)
• ipl_matches_2008_2024.csv (ball-by-ball data)

2. Run Complete Pipeline

python main.py

Choose option 1 for full setup: data processing → model training → predictions

3. Make Predictions

python predict_cli.py  # Traditional ML
python predict_with_nn.py  # Neural Network

📊 Model Performance

Model	Accuracy	ROC-AUC	Best For
Neural Network	73.85%	0.78	Overall best
Gradient Boosting	72.94%	0.82	Interpretability
XGBoost	68.81%	0.80	Feature importance
Random Forest	67.43%	0.75	Baseline

🔍 Key Prediction Features

1. Team venue performance (78% importance)
2. Season trends (9.4% importance)
3. Head-to-head records (4.5% importance)
4. Home ground advantage (4.2% importance)
5. Recent team form (3.3% importance)

💻 Usage Examples

Basic Prediction

from predict_match import predict_match

result = predict_match(
    team1="Chennai Super Kings",
    team2="Mumbai Indians", 
    venue="M.A. Chidambaram Stadium",
    toss_winner="Chennai Super Kings",
    toss_decision="bat"
)

print(f"🏆 Winner: {result['predicted_winner']}")
print(f"📈 Confidence: {result['win_probability']:.1f}%")

Interactive Menu

python main.py
# Choose from:
# 1. Complete pipeline
# 2. Data processing only  
# 3. Train models
# 4. Make predictions
# 5. Compare model performance

🛠️ Requirements

pip install pandas numpy scikit-learn xgboost matplotlib seaborn joblib

📋 Available Teams (2024)

• Chennai Super Kings
• Mumbai Indians
• Royal Challengers Bangalore
• Kolkata Knight Riders
• Delhi Capitals
• Sunrisers Hyderabad
• Punjab Kings
• Rajasthan Royals
• Gujarat Titans
• Lucknow Super Giants

🏟️ Supported Venues

• M.A. Chidambaram Stadium (Chennai)
• Wankhede Stadium (Mumbai)
• M. Chinnaswamy Stadium (Bangalore)
• Eden Gardens (Kolkata)
• Arun Jaitley Stadium (Delhi)
• And 5+ more major IPL venues

🎲 Prediction Output

Match Prediction Results
========================
Team 1: Chennai Super Kings
Team 2: Mumbai Indians
🏆 Predicted Winner: Chennai Super Kings
📊 Win Probability: 67.23%
🎯 Confidence: Medium

Team Probabilities:
CSK: 67.23% | MI: 32.77%

Match Factors:
🏠 Home Advantage: Yes (CSK)
📈 Recent Form: CSK (0.75) vs MI (0.60)
⚔️ Head-to-Head: CSK leads 52%
🎲 Toss Impact: Won toss, chose to bat

🚀 Future Enhancements

• Player impact analysis
• Weather condition integration
• Live match probability updates
• Web dashboard interface
• API endpoints for external use

---

Built with ❤️ for cricket analytics and machine learning enthusiasts