🏏 Pakistan Super League (PSL) — Match Winner Prediction

A machine learning project that predicts the winner of Pakistan Super League (PSL) matches using historical data. It includes data preprocessing, exploratory analysis, and model training using Random Forest , achieving high accuracy in predicting match outcomes.

---

📘 Project Overview

This project aims to predict the winner of Pakistan Super League (PSL) cricket matches using historical match data.
The model analyzes factors such as the playing teams, toss results, venues, and other match attributes to estimate the most likely winner.

This end-to-end solution includes:

• Data preprocessing and exploratory data analysis (EDA)
• Feature encoding and transformation
• Model training and evaluation using top-performing ML algorithm
• Visualization of performance metrics
• Model persistence for deployment and future use

---

🎯 Objective

The main goal of this project is to: Build a supervised machine learning model that can predict the winner of PSL matches based on available historical data.

Specifically, the project focuses on:

• Understanding the relationship between match features (teams, toss, venue, etc.) and outcomes.
• Training multiple ML model
• Evaluating model performance using metrics like accuracy, classification report, and confusion matrix.
• Saving trained pipelines for real-time or batch predictions.

---

🌟 Key Features

✅ Automated dataset discovery – Automatically locates and loads the PSL dataset from Kaggle environment.
✅ Robust preprocessing – Handles missing data, encodes categorical features, and scales numeric attributes.
✅ model trained – Random Forest Classifier
✅ Comprehensive Evaluation – Accuracy, classification report, and confusion matrix visualization.
✅ Label Encoding – Consistent numeric encoding of team names for compatibility with XGBoost.
✅ Prediction on Unseen Data – Make predictions for new/unseen match records.
✅ Model Persistence – Pipelines and encoders saved using joblib for reuse or deployment.

---

🧰 Tools & Technologies

Category	Tools/Packages
Language	Python 3.x
Libraries	pandas, numpy, matplotlib, scikit-learn, joblib
Environment	Kaggle / Jupyter Notebook
Visualization	matplotlib, ConfusionMatrixDisplay
Model Saving	joblib

---

📊 Dataset

Dataset Source: Pakistan Super League Dataset on Kaggle

Dataset Description:
The dataset includes detailed match-level information from PSL seasons, such as:

• team1, team2 — competing teams
• venue, city — match location details
• toss_winner, toss_decision — toss results
• result, winner — final match outcome

Target Variable:
winner — The team that won the match.

---

⚙️ Workflow

1. Import libraries and setup environment
2. Auto-detect and load dataset from /kaggle/input
3. EDA – Inspect missing values, value distributions, and target class balance
4. Feature Engineering – Encode categorical variables and scale numerical features
5. Train-Test Split (80-20)
6. Train two models – RandomForest
7. Evaluate models – Accuracy, classification report, and confusion matrix
8. Predict on unseen data
9. Save trained models and label encoder for future use

---

🧠 Model Training & Evaluation

Model Used:

• RandomForestClassifier

Both models were trained on preprocessed features using a unified scikit-learn Pipeline.

Evaluation Metrics:

• Accuracy

• Precision, Recall, F1-Score

• Confusion Matrix Visualization

  

Example Evaluation Output:

---

📚 References

• Scikit-Learn Documentation
• Kaggle PSL Dataset

---

👤 Author

Muqadas Ejaz

BS Computer Science (AI Specialization)

AI/ML Engineer

Data Science & Gen AI Enthusiast

📫 Connect with me on LinkedIn

🌐 GitHub: github.com/muqadasejaz

📬 Kaggle: Kaggle Profile

---

📎 License

This project is open-source and available under the MIT License.