Score Prediction in Indian Premier League through Data Mining
Shashwat Kumar Sinha
The IPL score is currently projected on the basis of run-rate and there is a lack of intuition on the basis of historical features as T20 cricket is considered unpredictable but through the advancement of data mining algorithms and as the Indian Premier League which has crossed the fifteen years landmark the ball-by-ball data is accessible and can be the subject of predictive analysis. The work we propose is based on historical features which affect the course of prediction quite drastically as the setting of the Indian Premier League has been in India for most of the seasons. A Linear Regression method has been proposed to predict the score of a IPL team on the basis of the current score given by the team and the wickets fallen after a certain interval ranging which impact a T20 match a lot and after the two strategic time-outs. The role of impact player is not considered since the rule has been implemented only for two years. A study has been proposed that how much these breaks affect the outcome of an IPL match and results in the prediction of an IPL match. The prediction accuracy has been increasing remarkably when the intervals of the games have been increased. The attributes considered affect the outcome at a much greater scale as compared to the other attribute and thus reduces the chances of over-fitting of the algorithm.
Under Process
Download Full Article
Quantam Computing Publication