Skip to content

This project uses supervised machine learning algorithms to build a model that predicts which IPL team has a higher probability of winning a match.

Notifications You must be signed in to change notification settings

SanjayKumhar/ipl-winning-probability-prediction

Repository files navigation

IPL Winning Probability Predictor

This project uses supervised machine learning algorithms to build a model that predicts which IPL team has a higher probability of winning a match. Although the Random Forest algorithm achieved a high accuracy of 99%, Logistic Regression was selected as the final model due to its practical applicability in real-world scenarios. Models with overly perfect results, such as 99% accuracy, often indicate overfitting or unrealistic predictions.

Features:

The model predicts the probable winning team using the various features from the dataset:

  • City: The city where the match is being played.
  • Batting Team: The team currently batting in the match.
  • Bowling Team: The team currently bowling in the match.
  • Current Run Rate: The rate at which the batting team is scoring runs per over.
  • Required Run Rate: The rate at which the batting team needs to score runs per over to win.
  • Other match-related features (e.g., match details, weather conditions, etc.)

Machine Learning Algorithms Used:

  • Random Forest
  • K-Nearest Neighbors (KNN) Classification
  • Decision Tree
  • Gradient Boost Classifier
  • Logistic Regression
  • AdaBoost Classifier
  • Voting Classifier

Key Findings:

  • Best Accuracy Achieved: Random Forest Algorithm (99% accuracy)
  • Selected Model: Logistic Regression, chosen for its simplicity and relevance in practical applications.

Dataset:

Releases

No releases published

Packages

No packages published