Evaluation of Machine Learning Techniques in Winner Predicting of The Hundred Games

Abstract

Cricket, being one of the most popular sports worldwide, attracted significant interest in developing accurate result prediction models. The Hundred is one of the several leagues that are contested in the world. It was important to research on accurate result prediction model in this league as the fan following and attention towards this league were increasing rapidly. The dataset was divided into training and test sets, and the models were evaluated on both datasets to measure their generalization performance. The findings demonstrated the potential of machine learning techniques in accurately forecasting Hundred match outcomes, enabling stakeholders to make informed decisions in the dynamic and unpredictable domain of cricket. Logistic Regression, Decision Tree and Random Forest models were implemented for match-winner prediction, while Gradient Boosting Regressor was used for score forecasting. This study gathered and analyzed Hundred data spanning multiple years, including player, match, team, and ball-to-ball information, to generate several conclusions that helped improve a team’s performance. The study highlights the applicability of supervised learning methods for enhancing decision-making in the dynamic and unpredictable environment of The Hundred.