Applications of modern classification techniques to predict the outcome of ODI Cricket

Abstract

Data mining and Machine learning in Sports analytics is a recent field in Computer Science. In this paper our goal is to predict the outcome of an ODI (One Day International) Cricket match. Outcome of an ODI Cricket match depends on several factors such as home game advantage, Day/Night, Toss, Innings (first or second), physical fitness of teams and dynamic strategies, a lot of which varies as the game proceeds. We have applied modern classification techniques -Naïve Bayesian, Support Vector Machines, and Random Forest, and conducted a comparative study based on their outcomes and performances. Based on the outcome of these models we have developed a tool COP (Cricket Outcome Predictor), which outputs the win/loss probability of an ODI match. The target audience of this tool involves teams playing cricket, and Sports Analysts in general.