Home

Welcome to the machineLearning wiki!

This project was made to predict the winner of matches in One Day International's in Cricket.

The final prediction done was on world cup 2015 with percentage of successful predicting ranging from 65% to 100% using different classification algorithms. Data of teams from year 2013 to 2015 before world cup started was taken. With factors like game played at home or outside, Batting run rate(runs scored per overs), Bowling Economy Rate(runs conceded in previous x games/over bowled), Batting Average(runs per wicket lost), Bowling Average(runs conceded by wickets taken), Batting Wicket Rate(wickets lost/balls batted), Bowling Strike rate(wickets taken/ balls bowled) for every game. The data of almost only 1.5 years was taken because the squad to play for world cup does not change much so close to the world cup. Average of these data was taken, so every team had an average value of all the above features. Now the algorithms are fed with the group matches data of the world cup and are tested to predict the matches like quarter finals, semi finals and final.

Also data is preprocessed for Batting run rate(runs scored per overs)- Bowling Economy Rate(runs conceded in previous x games/over bowled) Batting Average(runs per wicket lost)- Bowling Average(runs conceded by wickets taken) Batting Wicket Rate(wickets lost/balls batted)- Bowling Strike rate(wickets taken/ balls bowled) batting runrate* batting average - bowling economy rate *bowling average for every match and hence we can also predict who is going to win the match after the 1st innings has happened and 2nd innings is underway. More the game progresses into the 2nd innings the more correct are the percentages of predictions of the winners .