Journal of Sports Analytics | |
A machine learning approach to analyze ODI cricket predictors | |
article | |
Kalanka P. Jayalath1  | |
[1] Department of Mathematics and Statistics, University of Houston Clear Lake | |
关键词: Classification trees; cricket; logistic regression; ODI; regression trees; | |
DOI : 10.3233/JSA-17175 | |
来源: IOS Press | |
【 摘 要 】
As one-day international (ODI) games rise in popularity, it is important to understand the possible predictors that affect the game outcome. The home-field advantage, coin-toss result, bat-first or second, and day vs day-night game format are such popular variables being considered in the cricket literature. This article focuses on a comprehensive study of quantifying the significance of those important predictors via graphical ‘classification and regression tree’ (CART) and the popular logistic regression approaches. This study reveals the importance of the home-field advantage for major cricket playing nations in one-day international games but questions the uniformity of such factors under different playing conditions. Importantly, the home-field advantage is investigated further based on the opponent’s geographical location. Conclusively, the CART approach provides interesting and novel interpretations for popular predictors in ODI games.
【 授权许可】
Unknown
【 预 览 】
Files | Size | Format | View |
---|---|---|---|
RO202307140005006ZK.pdf | 213KB | download |