Abstract:
As the usage of statistical modelling became prevalent in sports field, the most fundamental problem appears to be predicting which of the teams wins the game. In this project, we try to explore a variety of supervised learning methods to forecast results of one of the most popular team games of the present age that is Cricket. We explore methods based on team compositions, team strengths, which are based on individual player’s performances in previous games as well as career performance, to predict the results for T20 cricket matches. We also try to select external features that are not dependent of individual players or teams to help achieve more accurate results
