Analysis of Factors Influencing Board Game Ownership Based on A Gradient Boosting Model

This study investigates the factors influencing board game ownership using a comprehensive dataset from BoardGameGeek. By applying statistical techniques such as correlation analysis, Lasso regression, Random Forest, and a Gradient Boosting Model (GBM), the paper explores how variables such as user ratings, complexity, playtime, age rating, and year of publication affect ownership. The analysis reveals that user engagement metricsparticularly the number of users who rated a gameare the strongest predictors of ownership. However, when visibility metrics are excluded, intrinsic attributes like game complexity, playtime, and accessibility (minimum age) emerge as significant drivers. The Gradient Boosting Model, trained with log-scaled ownership values, achieves a high score of 0.975 on the validation set, confirming its strong predictive performance. These findings provide actionable insights for developers, highlighting the importance of both marketing efforts and thoughtful design choices. The study contributes to the understanding of consumer behavior in the board game industry and offers a data-driven framework for optimizing game appeal and commercial success.