A machine learning project that predicts the probability of winning for IPL cricket matches during the second innings using various classification algorithms.

This project analyzes IPL match data to predict match outcomes based on real-time match situations. The model considers factors like current score, target, overs remaining, wickets fallen, and team performance to provide win probability predictions.

• Multiple ML Models: Logistic Regression, Decision Tree, Random Forest, SVM, XGBoost, and Neural Networks
• Real-time Predictions: Calculates win probability during second innings
• Interactive Web App: Streamlit-based user interface for live predictions
• Comprehensive Analysis: Detailed EDA and feature engineering
• Model Comparison: Performance evaluation across different algorithms

The project uses two main datasets:

• matches.csv: Contains match-level information (teams, venue, winner, etc.)
• deliveries.csv: Contains ball-by-ball data for each match

# Clone the repository
git clone https://github.com/[username]/ipl-match-predictor.git
cd ipl-match-predictor

# Install required packages
pip install pandas numpy seaborn matplotlib scikit-learn xgboost tensorflow streamlit

jupyter notebook MathcPredictor.ipynb

streamlit run app1.py

• runs_left: Runs required to win
• balls_left: Balls remaining in the innings
• wickets: Wickets remaining
• cur_run_rate: Current run rate
• req_run_rate: Required run rate
• batting_team: Team currently batting
• bowling_team: Team currently bowling
• city: Match venue

The Streamlit web app allows users to:

• Select batting and bowling teams
• Input current match situation (score, overs, wickets)
• Choose from multiple ML models
• Get real-time win probability predictions

• Team selection dropdown
• Real-time score input
• Model selection (Logistic Regression, SVM, Neural Network)
• Probability visualization

ipl-match-predictor/
│
├── MathcPredictor.ipynb    # Main analysis notebook
├── app1.py                 # Streamlit web application
├── README.md              # Project documentation
├── models/                # Saved model files
│   ├── logistic_model.pkl
│   ├── svc_model.pkl
│   ├── model_ann.h5
│   ├── scaler.pkl
│   └── encoded_columns.pkl
└── data/                  # Dataset files
    ├── matches.csv
    └── deliveries.csv

• Match Distribution: Analysis of matches played across different cities
• Team Performance: Win statistics for all IPL teams
• Score Distribution: First innings score patterns
• Feature Correlation: Relationship between features and match outcomes

The models show strong predictive capability with:

• High accuracy across different algorithms
• Good correlation between required run rate and match outcome
• Effective handling of categorical variables through one-hot encoding
• Proper scaling for distance-based algorithms

• IPL for providing the cricket data
• Scikit-learn and TensorFlow communities
• Streamlit for the web framework
• Cricket analytics community for inspiration

Note: This project is for educational and research purposes. Predictions are based on historical data and should not be used for betting or gambling purposes.