Hoop master: An integrated basketball simulation and coaching assistant system

ABSTRACT: Data-driven analytics are often underutilized in basketball coaching, leading to ineffective strategies and poor lineup decisions. To address this gap, the HoopMaster mobile Android application was developed, integrating advanced data analytics with interactive simulation to enhance coaching efficiency and game strategy. The system features an animated digital court board, successfully implemented through Unity Hub, allowing coaches to design, animate, and simulate basketball plays for improved communication and collaboration during practices and games. HoopMaster incorporates a Multiple Linear Regression (MLR) model to evaluate player performance metrics, such as points, assists, rebounds, and shooting percentages, and determine the optimal starting lineup. Using data from the Entertainment and Sports Programming Network, the model achieved an accuracy of 0.67, with a perfect recall of 1.00 and an F1-score of 0.67, ensuring high-impact player combination were consistently identified. Additionally, a Decision Three algorithm is used to recommend strategic plays tailored to each player’s strength and weaknesses. In testing across 60 play scenarios, the algorithm demonstrated 0.80 accuracy, with 1.00 precision for aggressive plays and 1.00 recall for conservative plays, achieving a macro-average F1-score of 0.76. By integrating MLR for intelligent lineup selection, decision trees for strategic paly recommendations, and animated simulation for visual clarity, HoopMaster delivers a holistic, data-driven approach to modern basketball coaching

Filipiniana