<![CDATA[This study explores the use of the Random Forest algorithm to predict the popularity of Roblox games based on key player engagement metrics such as Active players, Likes, Dislikes, Favourites, and Rating. Using a dataset of the top 1000 games on Roblox, the model was trained to predict the total number of Visits for each game, serving as the target variable. The model was evaluated using multiple metrics, including Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-squared (R²), with an R² score of 0.7814, indicating a strong ability to predict game popularity. Key findings include the significant role of Dislikes in determining game success, which had the highest importance score in the model, followed by Likes and Active players. These insights suggest that negative feedback, as captured by Dislikes, plays an important role in shaping a game's visibility and success, alongside positive engagement metrics. Despite the promising results, the study acknowledges limitations such as reliance on publicly available data and the potential for data sparsity in less popular games. The study contributes to the understanding of metaverse game dynamics, specifically on platforms like Roblox, by providing a robust predictive model that can aid game developers in optimizing their games for better player engagement and long-term success. Future research directions include incorporating additional player behavior data and testing alternative machine learning models to further enhance predictive accuracy and address the limitations of this study.]]>
Copyrights © 2025
