Cricket Analytics: ODI Match Score Prediction and Resource Metric Modeling using Machine Learning Model

Sports analytics using machine learning is no longer a young area of computer science research that presents several difficulties. The purpose of this study is to develop a method for predicting game outcomes during an ODI cricket competition. Like all other papers, we adopted the factor-based prediction meaning the prediction will depend on various game factors like wickets left, overs left, current run, run rate, etc. Sometimes the models can not make good predictions based on these factors and we tried to improve that by adding weight to overs and wickets because these two vastly control the score of the game. We named this feature the resource factor. We used linear regression, polynomial regression, elastic net regression, Xgboost, and random forest and found that random forest works best for our dataset. For more interaction, we used a web application and we applied our best model to predict the score. For the test phase, the R² score for the random forest regressor is 0.9899 which is pretty good, and for the MAE and MSE scores, we have got 3.3569 and 37.3555.