Lists
Lists
Sharma, Pulkit, Hassan, Bilal and Wasiq, Muhammad Farooq (2024) AutoCoach: an automated playing eleven selection framework in cricket. In: 2024 International Conference on Emerging Techniques in Computational Intelligence (ICETCI), 22-24 August 2024, Mahindra University, Hyderabad (India).
Cricket is one of the famous international sports where both teams select their best eleven against each other from the available pool of players usually a total size of 15 to 16. The team selection process is usually performed by the coaches using their observation and consultations etc. In simple words, by looking at the possible combinations of opposite sides, the best possible combinations are suggested. In our work, we tried to automate this process using machine learning models like Support Vector Regressor, Linear Regression Random Forest regressor etc. To train and test the model, the data was crawled from the Indian Premier League (IPL), which follows 20 over format and data for more than 10 seasons is used. More importantly, multiple related features for batters and bowlers were accumulated to develop all possible combinations for each playing team in a single match. In total, we selected 4 different teams and developed three combinations for each. Then, in a single match between two teams, the combinations were compared based upon accumulative score e.g., matching scores for different combinations are compared with ground truth. In our opinion, this is one of the unique contributions made for automated team selection in cricket specific to the format and encompassing performance indicators not only accumulated from IPL but also from the International Cricket Council (ICC).
Available under License Creative Commons Attribution 4.0.
Download (1MB) | Preview
 |
View Item |