Research Information System
14th International Conference on Innovations in Bio-Inspired Computing and Applications and 13th World Congress on Information and Communication Technologies, IBICA-WICT 2023, Kochi, India, 14 - 15 December 2023, vol.1229 LNNS, pp.144-153, (Full Text)
The sports industry has witnessed significant growth, capturing the attention of a diverse audience. Within this burgeoning industry, athletes hold a substantial share, and their performances and ratings assume paramount significance, particularly in the determination of their remuneration. This study delves into the realm of sports video games, focusing on the NBA 2K20 official video game, employing regression analysis and various Machine Learning (ML) algorithms, including k-Nearest Neighbors (Knn), Lasso, and Extreme Gradient Boosting (XGBoost), to evaluate player ratings. Moreover, the study investigates the impact of ensemble methods on model performance. In this study, we applied ML algorithms within the domain of sports video games. This endeavour not only yielded valuable insights into the gaming experience but also highlighted the potential for data-driven decision-making within the sports industry. This study shows that Knn consistently exhibits superior predictive accuracy in both paradigms. While Boosting showed improvement, Bagging emerged as particularly prominent, engendering incremental enhancements in MAE and R2 values across all models, highlighting its effectiveness in ameliorating predictive precision.