Prediction of IPL Player Selection from T20 formats using a Novel Data Balancing Approach

Selecting players in the squad for the very popular Indian Premier League (IPL) is crucial. It is influenced by various factors, like popularity, fitness, and the players' performance in various cricket formats. This research paper focuses on creating a model to check whether an Indian player will be selected for IPL by analyzing the data from the year 2021-2023 across various T20 cricket tournaments, namely Sayed Mushtaq Ali Trophy, Irani Trophy, and Duleep Trophy. The collected data was preprocessed and consolidated into two datasets: batsmen and bowlers with unique player entries separately. The dataset was biased towards the non-ipl players. Thus, to address the issue of an imbalanced dataset, we tried various data balancing techniques and introduced a novel, problem-specific approach to balancing the dataset. All these balanced datasets were trained on a logistic regression model to predict player selection. Our results show that the problem-specific data balancing technique outperformed the state-of-the-art techniques, offering a more accurate and reliable solution for IPL player selection. This paper contributes a new data balancing strategy and provides valuable insights for improving player selection models in sports analytics.