Recommender Selection System for the Game Using Bonferroni Mean Based TOPSIS

In this paper, a recommendation system is applied to the motor selection menu in the game “Blar Rising”. This recommendation feature is provided in the selection menu to replace the commonly used randomize feature in many games. The recommendation feature takes into account each criterion of the available motors. The determination of weights is done using the ROC method, which considers the level of importance of the criteria. Therefore, in the game, there is a selection of tracks used to differentiate the importance level for weight determination. From the results obtained, in this case, both the TOPSIS and TOPSIS-BM methods exhibit good performance in making decisions. In its implementation, Bonferroni Mean (BM) is used to generalize the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method. This can be observed from the difference in the results for each alternative. The TOPSIS-BM method yields similar results for each alternative, while the original TOPSIS method shows significant differences. Generalization makes the results more stable and capable of handling situations where criteria have extreme values. Usability testing across five aspects yielded an overall score of 3.43, with individual ratings for learnability (2.95), efficiency (3.87), memorability (3.67), error (3.20), and satisfaction (3.45). This suggests favorable usability acceptance among players for the recommendation feature.