One Day International Cricket Match Score Prediction Using Machine Learning Approaches

Cricket is the game that binds Bangladeshis to-gether. The datasets were split into different parts based on the time of the cricket matches to experiment with the results. We have used Random Forest Regression and Linear Regression to experiment with the results. We have worked on different sets of independent variables while working with different datasets but runs and wickets were constant features. The linear Regression technique gave 0.58 of r-squared metric and its MAE: 32.857046003557706, MSE: 1793.4360657965133, RMSE: 42.34897951304746 values indicate that the regression model is making predictions that are reasonably close to the actual ob-served values. However, there is still some room for improvement. The r-squared metric of 0.787 was derived from a relatively brief dataset, MAE: 20.48880467759807, MSE: 690.1167559064677, RMSE: 26.270073389818837 metrics indicate that the regression model is making predictions that are very close to the actual observed values. Random Forest Regression's r-squared metric resulted in 0.807 but the learning curve pointed out an overfit model. Thus, linear regression contained the best result. We have applied the regularisation technique to achieve a better result. The dataset was collected from Github and Kaggle, sizes are (350899, 15) and (4037, 9) respectively.