Data Analysis of Cricket Score Prediction

Abstract

Cricket, a game played on 22-yard strip among 22 players with a bat and ball, is one among the most popular sport in the world. Even though it is played by a lesser number of countries, the sport is followed all across the globe. People are not only interested in following this sport, they also try predicting the flow of the match. Predicting the flow of a cricket match has always been a strenuous task as a particular player may not perform the same way against every opposition nor will his performance be the same in every venue. Also, the way a player performs depends according to the dynamics of the game. While predicting the flow of the match, a majority of the people tend to give more weightage to the previous results. In this paper, we have come up with a model which not only takes previous results into consideration but also the opposition, the venue and the current state of the match such as, number of wickets fallen, number of overs remaining, the way the players have fared till that moment. We have developed various algorithms to find batting and bowling index of the players involved in the match. These indices, along with a special feature, RunFactor form the input parameters to our machine learning model. The generated output from this model is the number of runs that will be scored in the particular over. Compiling these, we estimate the final score scored by that team in a One Day International (ODI).