Prediction of Player Price in IPL Auction Using Machine Learning Regression Algorithms

Abstract

In this work, we have applied machine learning-based algorithms that predicts the cost at which a player can be sold in the Indian Premier League Auction. We estimated the players' selling price using their past performance parameters like runs, balls, innings, wickets and matches played. Tests were carried out in various machine learning models like Decision Tree Regressor, K-Nearest Neighbors (KNN), Linear Regression, Stochastic Logistic Regression, Random Forest Regressor and Support Vector Regression (SVR). Among these SVR and Linear Regression gave best results for predicting batsman and bowlers respectively. These algorithms can produce fast and accurate results within 3 seconds, helping auctioneers make quick decisions. We have also considered inflation factor and mapping of the same to the budget during the training of the model.